CN116151472A - Electric automobile short-term energy consumption prediction method based on Internet of vehicles data mining - Google Patents
Electric automobile short-term energy consumption prediction method based on Internet of vehicles data mining Download PDFInfo
- Publication number
- CN116151472A CN116151472A CN202310239797.XA CN202310239797A CN116151472A CN 116151472 A CN116151472 A CN 116151472A CN 202310239797 A CN202310239797 A CN 202310239797A CN 116151472 A CN116151472 A CN 116151472A
- Authority
- CN
- China
- Prior art keywords
- energy consumption
- mileage
- fragment
- predicted value
- measured
- 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.)
- Pending
Links
- 238000005265 energy consumption Methods 0.000 title claims abstract description 128
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000007418 data mining Methods 0.000 title claims abstract description 15
- 239000012634 fragment Substances 0.000 claims abstract description 101
- 230000008569 process Effects 0.000 claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims abstract description 10
- 238000007599 discharging Methods 0.000 claims abstract description 9
- 238000003062 neural network model Methods 0.000 claims description 23
- 239000013598 vector Substances 0.000 claims description 17
- 230000007787 long-term memory Effects 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 9
- 230000015654 memory Effects 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 5
- 238000003066 decision tree Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 2
- 238000012549 training Methods 0.000 description 20
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 208000019901 Anxiety disease Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000036506 anxiety Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- General Health & Medical Sciences (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the field of energy management of electric vehicles, and relates to an electric vehicle short-term energy consumption prediction method based on internet of vehicles data mining; the method comprises the steps of obtaining operation data and working condition data of the electric automobile from an Internet of vehicles data platform; dividing the operation data into mileage fragments according to the discharging process; extracting characteristic variables related to the energy consumption of the electric automobile from the working condition data according to the related coefficients; parameter calculation is carried out on the selected characteristic variables in the divided mileage fragments, and input variables of mileage fragments to be measured are constructed; inputting the input variable of the mileage fragment to be measured into the trained integrated learning model to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured; and calculating the predicted value of the energy consumption of the electric automobile in the mileage fragment to be measured by using the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured. The electric automobile energy consumption prediction model designed based on the energy consumption adjustment factors can more accurately predict the energy consumption level of the electric automobile.
Description
Technical Field
The invention relates to the field of energy management of electric vehicles, mainly relates to the problem of prediction of the energy consumption of the whole vehicle in a period of time in the future, and in particular relates to a short-term energy consumption prediction method of the electric vehicle based on internet of vehicles data mining.
Background
Compared with other new energy vehicles, the electric vehicle has obvious advantages, and the related technology of the electric vehicle is the focus of research in the current automobile industry and academia due to the running economy and environmental friendliness. Many key technologies have not achieved breakthroughs including on-board battery energy management. Electric automobile energy consumption prediction is one of the key works.
The current electric automobile energy consumption prediction method is mostly expressed as follows: building a common neural network framework, training a model by using the energy consumption value and the characteristic variable related to the energy consumption, and finally outputting an energy consumption predicted value; and combining the urban road network structure where the vehicle is located and traffic conditions, and carrying out energy consumption prediction and the like in a green wave band scene. These traditional methods are relatively limited to specific vehicle models in specific environments and conditions, and the trained models are not necessarily applicable to other vehicles under other working conditions; in the real driving working condition, the abrupt environmental condition and the special traffic condition are difficult to predict, and the energy consumption of the whole vehicle in a longer period in the future is also difficult to predict accurately, so that the prediction result does not have high credibility and insufficient generalization capability.
The method has the advantages that the energy consumption of the electric automobile in the future period is predicted, the running condition of the future automobile can be reflected, real-time driving mileage information and driving decisions are provided for users, the use confidence of the users on the electric automobile is improved, and the mileage anxiety problem is relieved. The existing automobile energy consumption prediction method mainly comprises the following steps: and constructing a multidimensional matrix based on the historical data of the energy consumption related characteristic variables, and obtaining a final energy consumption predicted value through training the input variables and the output variables. In fact, the energy consumption level of the electric automobile is not only related to the whole automobile working condition (such as speed, historical whole automobile energy consumption value, battery system parameters and the like) in real time, but also complex conditions such as road wet and slippery conditions in rainy and snowy weather, steep road gradient and the like and traffic working conditions (such as urban morning and evening peaks, workday double holidays and the like) are not ignored. The data mining method reflects possible future situations according to the historical working conditions, and can obtain accurate energy consumption values under the future driving conditions under the condition that the input features are selected correctly and the data processing is proper. However, the characteristic information influencing factors related to energy consumption are complex, the acquisition is difficult, and modeling work is difficult; when online prediction of energy consumption is performed, the input features required by the prediction model cannot guarantee real-time performance and accuracy, and the generated prediction effect may be unsatisfactory.
Disclosure of Invention
Based on the problems existing in the prior art, the invention provides an electric vehicle short-term energy consumption prediction method based on Internet of vehicles big data mining, which comprises the following steps:
acquiring operation data and working condition data of the electric automobile from the Internet of vehicles data platform, and preprocessing and normalizing the data;
dividing the operation data into a charging process and a discharging process, and dividing mileage fragments according to the discharging process;
extracting characteristic variables related to the energy consumption of the electric automobile from the working condition data according to the related coefficients;
in each divided mileage fragment, carrying out parameter calculation on the selected characteristic variable to construct an input variable of the mileage fragment to be measured;
inputting the input variable of the mileage fragment to be measured into the trained integrated learning model to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured;
and calculating the predicted value of the energy consumption of the electric automobile in the mileage fragment to be measured by using the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured.
The invention has the beneficial effects that:
the patent of the invention proposes the concept of an energy consumption adjustment factor, which is used for representing the coupling relation between the historical energy consumption and the future energy consumption, and aims to obtain the predicted value of the future energy consumption by combining the known historical energy consumption and the energy consumption factor. The energy consumption adjustment factor and the characteristic variable related to the energy consumption have a certain linear and nonlinear relation, so that the determination of the specific input and output characteristic variable type and the true value thereof and the correct obtaining of the relation between the specific input and output characteristic variable type and the true value thereof are key points for accurately predicting the energy consumption. Therefore, a prediction model frame is designed based on the energy consumption adjustment factors, and the obtained electric vehicle energy consumption prediction model can more accurately predict the energy consumption level of the electric vehicle; the result can ensure that the device has certain real-time performance, can obtain a reasonable and determined energy consumption value before the start of a subsequent mileage, and provides effective input for a driver to make driving decisions.
Drawings
Fig. 1 is a flowchart of a method for predicting short-term energy consumption of an electric vehicle according to an embodiment of the present invention;
FIG. 2 is a block diagram of an ensemble learning model training predictive framework in accordance with an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Fig. 1 is a flowchart of a method for predicting short-term energy consumption of an electric vehicle according to an embodiment of the present invention; as shown in fig. 1, the invention provides an electric vehicle short-term energy consumption prediction method based on internet of vehicles data mining, which comprises the following steps:
101. acquiring operation data and working condition data of the electric automobile from the Internet of vehicles data platform, and preprocessing and normalizing the data;
in the embodiment of the invention, the running data of the automobile in a period of time or a period of driving mileage is required to be acquired from the electric automobile networking data platform. To obtain more accurate prediction results, the operation data needs to be preprocessed, which includes: checking whether the data accords with the time sequence, checking the data lack and abnormal conditions, and the like. The purpose of the preprocessing is to minimize the impact of objective errors on the results.
In the embodiment of the invention, the working condition data comprise historical whole vehicle working conditions, road working conditions and traffic working conditions. The whole vehicle working condition comprises a vehicle speed, a historical whole vehicle energy consumption value, battery system parameters and the like, the road working condition comprises a road wet slip condition, a road gradient steep degree and the like in rainy and snowy weather, and the traffic working condition comprises a vehicle traffic volume and the like in the conditions of urban morning and evening peaks, workdays and double holidays and the like.
102. Dividing the operation data into a charging process and a discharging process, and dividing mileage fragments according to the discharging process;
in the embodiment of the invention, the normalized operation data is required to be segmented according to a certain rule. Firstly, dividing a discharging process and a charging process, and omitting all vehicle charging processes because the final output result is energy consumption. Aiming at a discharging process, dividing mileage fragments by taking every s kilometers as a reference, and finally dividing n mileage fragments totally for the embodiment of the invention.
103. Extracting characteristic variables related to the energy consumption of the electric automobile from the working condition data according to the related coefficients;
in the embodiment of the invention, the characteristic variables related to the energy consumption of the electric automobile are extracted from the historical whole vehicle working conditions, road working conditions and traffic working conditions by combining the prior art with the common knowledge of physics. The linear correlation between different variables is judged by using a correlation coefficient such as a pearson correlation coefficient or a spearman correlation coefficient, and the variable with higher correlation with the energy consumption value is selected to be used as an input variable for the modeling analysis.
104. In each divided mileage fragment, carrying out parameter calculation on the selected characteristic variable to construct an input variable of the mileage fragment to be measured;
in the embodiment of the invention, parameter calculation is carried out on the selected characteristic variable in each divided mileage fragment, and finally the required input variable is constructed. For example: the segment average speed, average energy consumption per kilometer, average battery SOC loss, battery temperature extremum, average road traffic flow and congestion level, average road grade, etc. are calculated, which can directly reflect the energy consumption level in the segment in value.
Because of the difference of the dimensions of different feature variables, normalization processing is needed for all feature vectors to eliminate the influence of different variables on data and ensure efficient calculation. The normalized results are scaled to be within 0,1, and the calculation formula is shown as formula (1):
wherein x is an input variable, x min and xmax Respectively, the minimum and maximum values of the input variable.
105. Inputting the input variable of the mileage fragment to be measured into the trained integrated learning model to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured;
in the embodiment of the invention, in the training process, an energy consumption adjustment factor y (x) of a certain mileage fragment needs to be introduced to represent the ratio of the average energy consumption value per kilometer of the next mileage fragment to the mileage fragment, and the definition formula is shown as formula (2):
in the formula ,ΔQj+1 and ΔQj Represents the average energy consumption per kilometer of the j+1st and j-th fragments, j E [1, n ]]. The energy consumption adjustment factor y (x) will be the output variable of the model.
In the training process of the embodiment, the input vector represented by the formula (1) and the output vector represented by the formula (2) are respectively used as the input and output of the energy consumption prediction model, and the relation between the input variable of the mileage fragment to be tested and the predicted value of the energy consumption adjustment factor of the mileage fragment to be tested is learned through the linear and nonlinear relation between the training input and output, wherein the input and output of the model in the training process are shown in the formula (3):
due to the average energy consumption value delta Q per kilometer of the n+1th mileage fragment n+1 Is the predicted value needed to be obtained later, belongs to the unknown quantity, so the history data does not comprise the energy consumption adjustment factor y of the nth segment n (x)。
Through the iterative training, in the embodiment of the invention, the linear and nonlinear relations between the input and output variables in the historical data can be obtained through model training, and integrated learning models are respectively built, wherein the integrated learning models comprise an LSTM long-short-term memory neural network and an XGBoost model. Processing input variables of the mileage fragment to be measured by adopting the trained LSTM long-term memory neural network model and the trained XGBoost model respectively to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured under the LSTM long-term memory neural network model and the XGBoost network model; adopting an RBF radial basis function neural network model to adaptively correct the predicted value of the energy consumption adjustment factor output by the LSTM long-term memory neural network model and the XGBoost network model; and weighting to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured.
Considering that the LSTM long-short-term memory neural network model is more suitable for predicting long data, and the XGBoost network model is more suitable for predicting short data, the invention divides the input vector X of the mileage fragment to be measured into a plurality of unit data fragments with the length of P and a plurality of unit data fragments with the length of Q respectively by each mileage fragment, so that the two network models are respectively predicted according to the good data length of the two network models; inputting the P unit data fragments into a trained LSTM long-short-term memory neural network model, and calculating the variance of a radial basis function according to the length of each unit data fragment and the length of a target data fragment; performing data fitting according to the P unit data fragments, and outputting to obtain a first energy consumption adjustment factor predicted value of the mileage fragment to be measured; inputting the Q unit data fragments into a trained XGBoost network model, and performing data fitting according to the Q unit data fragments; outputting a second energy consumption adjustment factor predicted value of the mileage fragment to be measured; and carrying out weighted summation on the predicted value of the first energy consumption adjusting factor and the predicted value of the second energy consumption adjusting factor to obtain the predicted value of the energy consumption adjusting factor of the final mileage fragment to be measured, wherein P is more than Q, and the sizes of P and Q can be adjusted according to the predicted effect during training.
For the LSTM long-term memory neural network model, the length of each unit data segment is u=x/P, and for the XGBoost network model, the length of each unit data segment is v=x/Q; meanwhile, considering that the data length has a certain influence on the model prediction result, the existing network model does not consider the influence of the data length on the prediction result, and under some conditions, the problems of data overfitting, sample blurring and the like can occur, so that the network model is difficult to converge, and the situations of data misjudgment and the like can occur; the method also adaptively adjusts the variance of the RBF function according to the length of the unit data segment and the length of the target data segment, predicts according to the data length adapted by the network model, solves the problems of fuzzy sample and data overfitting, and greatly improves the detection accuracy.
And substituting the input variable and the output variable of the n-1 mileage fragments into two network models respectively, and dividing a training set and a verification set for training and verification respectively. The linear regression relationship between the finally obtained output variable and the input variable is shown in the following formula (4) and formula (5):
wherein ,
a first energy consumption adjustment factor predicted value of a mileage fragment to be measured, which is output by the LSTM long-term memory neural network model, is represented; sigma represents an activation function, X P Input vector W representing mileage fragment to be tested input by LSTM network model input layer xo A weight matrix representing the input layer and the output layer; h P-1 A p-th unit data segment vector representing the hidden layer output; w (W) ho A weight matrix representing the hidden layer and the output layer; b o1 Representing a bias constant of the output layer; w (w) i The i weight of the input layer and the output layer is represented, and L represents the weight dimension of the input layer and the output layer; x is x p Representing the p-th unit data segment vector, x, input by the input layer i Representing a target unit data segment vector; />
And the variance of the unit data segment length and the target data segment length of the LSTM long-term memory neural network model is represented. />
A second energy consumption adjustment factor predicted value of the mileage fragment to be measured, which is output by the XGBoost network model, is represented; superscript (j) represents the j-th decision tree; x is X Q Input vector x representing mileage fragment to be tested input by XGBoost network model input layer q A q-th unit data segment vector representing an input of the input layer; f (x) q ) Representing the result of the j-th decision tree prediction.
Then the radial basis function neural network structure surrounding the RBF is utilized, and the training output based on the LSTM model and the XGBoost model is realized by utilizing the strong local nonlinear mapping capability
And a real energy consumption adjustment factor Y, so that a strong learner integrating LSTM-XGBoost and RBF is obtained through training. The RBF network model has the function of taking the real energy consumption adjustment factor Y as a reference, carrying out weighted fusion on the predicted values of the energy consumption adjustment factors output by the LSTM model and the XGBoost model, and achieving the purpose of correction or correction so as to ensure that the model effect is optimal. The resulting energy consumption adjustment factor->
The method comprises the following steps:
wherein ,
representing the predicted value of the energy consumption adjustment factor of the nth mileage fragment; omega 1 Weight coefficient representing LSTM long-term memory neural network model, +.>
First energy consumption adjustment factor predictive value omega of mileage segment to be measured output by LSTM long-term memory neural network model 2 Weight coefficient representing XGBoost network model, < ->
And the second energy consumption adjustment factor predicted value of the mileage fragment to be measured, which is output by the XGBoost network model, is represented.
After the construction of all models is completed, all input variables X of the nth mileage fragment are calculated n ＝(x 1,n x 2,n x 3,n … x k,n ) Substituting the energy consumption adjustment factor prediction value into the formulas (4), (5) and (6) to obtain the energy consumption adjustment factor prediction value of the nth mileage fragment
Then combining the formula (7), the average energy consumption predictive value per kilometer of a future fragment, namely the n+1th fragment, can be obtained>
Similarly, after the electric automobile runs through the n+1th mileage fragment, the values X of all the input features of the n+1th fragment can be acquired in real time n ＝(x 1,n+1 x 2,n+1 x 3,n+1 … x k,n+1 ) The energy consumption adjustment factor of the segment can be obtained in the same way as in the formula (6)
And calculating the average energy consumption per kilometer predicted value of the n+2-th segment by using the formula (7)>
Through the above process, the average energy consumption value per kilometer in the next trip after each mileage segment can always be obtained in advance. And after the segment is finished subsequently, obtaining a real energy consumption value, comparing the real energy consumption value with a predicted value in real time, setting a predicted fault-tolerant interval, determining whether the predicted value or a true value is substituted into a previously established model according to whether the predicted value is in the fault-tolerant interval, and repeating training and updating.
106. And calculating the predicted value of the energy consumption of the electric automobile in the mileage fragment to be measured by using the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured.
In the real driving situation, the driving decision can be adjusted in real time according to the predicted value, and the aim of optimizing energy management is achieved by timely controlling the speed of the vehicle and the running of load equipment in the vehicle.
FIG. 2 is a schematic diagram of training and prediction, recording the steps and methods of training, prediction and analysis after model creationA method of manufacturing the same. Assuming that n mileage fragments exist, firstly performing offline training of a model in n-1 mileage fragments, and obtaining a regression relationship after training and verification as follows
Assuming that the machine learning model adopted at the moment is built, substituting the nth mileage fragment into a regression relation to obtain an output variable predicted value +.>
Obtaining an energy consumption prediction value by definition calculation>
When the nth mileage fragment is finished by the automobile, obtaining the true energy consumption value delta Q n+1 . A prediction fault-tolerant interval can be set, whether the predicted value is in the fault-tolerant interval is judged after the actual value is obtained, and the actual value or the predicted value is substituted into the model for updating training according to judgment decision.
Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be implemented by a program to instruct related hardware, the program may be stored in a computer readable storage medium, and the storage medium may include: ROM, RAM, magnetic or optical disks, etc.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (8)
1. The short-term energy consumption prediction method for the electric automobile based on the Internet of vehicles data mining is characterized by comprising the following steps:
acquiring operation data and working condition data of the electric automobile from the Internet of vehicles data platform, and preprocessing and normalizing the data;
dividing the operation data into a charging process and a discharging process, and dividing mileage fragments according to the discharging process;
extracting characteristic variables related to the energy consumption of the electric automobile from the working condition data according to the related coefficients;
in each divided mileage fragment, carrying out parameter calculation on the selected characteristic variable to construct an input variable of the mileage fragment to be measured;
inputting the input variable of the mileage fragment to be measured into the trained integrated learning model to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured;
and calculating the predicted value of the energy consumption of the electric automobile in the mileage fragment to be measured by using the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured.
2. The method for predicting short-term energy consumption of an electric vehicle based on internet of vehicles data mining according to claim 1, wherein the working condition data comprises historical whole vehicle working conditions, road working conditions and traffic working conditions.
3. The method for predicting the short-term energy consumption of the electric automobile based on the internet of vehicles data mining according to claim 1, wherein the step of inputting the input variable of the mileage fragment to be tested into the trained integrated learning model to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be tested comprises the steps of respectively adopting a trained LSTM long-short-term memory neural network model and a trained XGBoost network model to process the input variable of the mileage fragment to be tested to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be tested under the LSTM long-term memory neural network model and the XGBoost network model; adopting an RBF radial basis function neural network model to adaptively correct the predicted value of the energy consumption adjustment factor output by the LSTM long-term memory neural network model and the XGBoost network model; and weighting to obtain the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured.
4. The method for predicting the short-term energy consumption of the electric automobile based on the internet of vehicles data mining according to claim 1, wherein the method for adaptively correcting the predicted value of the energy consumption adjustment factor output by the LSTM long-short-term memory neural network model and the XGBoost network model by adopting the RBF radial basis neural network model comprises the steps of respectively dividing an input vector of a mileage fragment to be detected into a plurality of unit data fragments with the length of P and a plurality of unit data fragments with the length of Q; inputting the P unit data fragments into a trained LSTM long-short-term memory neural network model, and calculating the variance of a radial basis function according to the length of each unit data fragment and the length of a target data fragment; performing data fitting according to the P unit data fragments, and outputting a first energy consumption adjustment factor predicted value of the estimated mileage fragment to be measured; inputting the Q unit data fragments into a trained XGBoost network model, and performing data fitting according to the Q unit data fragments; outputting a second energy consumption adjustment factor predicted value of the estimated mileage fragment to be measured; and carrying out weighted summation on the predicted value of the first energy consumption adjusting factor and the predicted value of the second energy consumption adjusting factor to obtain the predicted value of the energy consumption adjusting factor of the final mileage fragment to be measured, wherein P is more than Q.
5. The short-term energy consumption prediction method for an electric vehicle based on internet of vehicles data mining according to claim 4, wherein a calculation formula of a predicted value of a first energy consumption adjustment factor of a mileage fragment to be measured output by an LSTM long-term memory neural network model is expressed as follows:
wherein ,
a first energy consumption adjustment factor predicted value of a mileage fragment to be measured, which is output by the LSTM long-term memory neural network model, is represented; sigma represents the activation function，X P Input vector W representing mileage fragment to be tested input by LSTM network model input layer xo A weight matrix representing the input layer and the output layer; h P-1 A p-th unit data segment vector representing the hidden layer output; w (W) ho A weight matrix representing the hidden layer and the output layer; b o1 Representing a bias constant of the output layer; w (w) i The i weight of the input layer and the output layer is represented, and L represents the weight dimension of the input layer and the output layer; x is x p Representing the p-th unit data segment vector, x, input by the input layer i Representing a target unit data segment vector; />
And the variance of the unit data segment length and the target data segment length of the LSTM long-term memory neural network model is represented.
6. The short-term energy consumption prediction method for an electric vehicle based on internet of vehicles data mining according to claim 4, wherein a calculation formula of a predicted value of a second energy consumption adjustment factor of a mileage fragment to be measured output by an XGBoost network model is expressed as:
wherein ,
a second energy consumption adjustment factor predicted value of the mileage fragment to be measured, which is output by the XGBoost network model, is represented; superscript (j) represents the j-th decision tree; x is X Q Input vector x representing mileage fragment to be tested input by XGBoost network model input layer q A q-th unit data segment vector representing an input of the input layer; f (x) q ) Representing the result of the j-th decision tree prediction.
7. The short-term energy consumption prediction method of an electric automobile based on internet of vehicles data mining according to claim 1, wherein the formula of the predicted value of the energy consumption adjustment factor of the mileage fragment to be measured is expressed as:
wherein ,
representing the predicted value of the energy consumption adjustment factor of the nth mileage fragment; omega 1 Weight coefficient representing LSTM long-term memory neural network model, +.>
First energy consumption adjustment factor predictive value omega of mileage segment to be measured output by LSTM long-term memory neural network model 2 Weight coefficient representing XGBoost network model, < ->
And the second energy consumption adjustment factor predicted value of the mileage fragment to be measured, which is output by the XGBoost network model, is represented.
8. The short-term energy consumption prediction method of the electric automobile based on the internet of vehicles data mining according to claim 1, wherein the formula for calculating the energy consumption prediction value of the electric automobile in the mileage fragment to be measured is expressed as follows:
wherein ,
the energy consumption predicted value of the (n+1) th mileage fragment to be measured is represented; />
Representing the predicted value of the energy consumption adjustment factor of the nth mileage fragment; ΔQ n Representing the energy consumption value of the nth mileage fragment. />
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310239797.XA CN116151472A (en) | 2023-03-13 | 2023-03-13 | Electric automobile short-term energy consumption prediction method based on Internet of vehicles data mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310239797.XA CN116151472A (en) | 2023-03-13 | 2023-03-13 | Electric automobile short-term energy consumption prediction method based on Internet of vehicles data mining |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116151472A true CN116151472A (en) | 2023-05-23 |
Family
ID=86358293
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310239797.XA Pending CN116151472A (en) | 2023-03-13 | 2023-03-13 | Electric automobile short-term energy consumption prediction method based on Internet of vehicles data mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116151472A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116754828A (en) * | 2023-08-21 | 2023-09-15 | 济南瑞源智能城市开发有限公司 | Intelligent tunnel energy consumption monitoring method, device and medium |
-
2023
- 2023-03-13 CN CN202310239797.XA patent/CN116151472A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116754828A (en) * | 2023-08-21 | 2023-09-15 | 济南瑞源智能城市开发有限公司 | Intelligent tunnel energy consumption monitoring method, device and medium |
| CN116754828B (en) * | 2023-08-21 | 2023-12-01 | 济南瑞源智能城市开发有限公司 | Intelligent tunnel energy consumption monitoring method, device and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111832225B (en) | Method for constructing driving condition of automobile | |
| CN111301426B (en) | Method for predicting energy consumption in future driving process based on GRU network model | |
| CN108556682B (en) | Driving range prediction method, device and equipment | |
| CN110910531B (en) | Rapid pavement friction coefficient detection method based on vehicle-mounted OBD information | |
| CN111339712A (en) | Method for predicting residual life of proton exchange membrane fuel cell | |
| CN114664091A (en) | Early warning method and system based on holiday traffic prediction algorithm | |
| CN115271186B (en) | Reservoir water level prediction and early warning method based on delay factor and PSO RNN Attention model | |
| CN112966871A (en) | Traffic jam prediction method and system based on convolution long-short term memory neural network | |
| CN116151472A (en) | Electric automobile short-term energy consumption prediction method based on Internet of vehicles data mining | |
| CN112860782A (en) | Pure electric vehicle driving range estimation method based on big data analysis | |
| CN116523177A (en) | Vehicle energy consumption prediction method and device integrating mechanism and deep learning model | |
| CN114855570A (en) | Municipal road maintenance strategy processing method and device and computer equipment | |
| CN111967308A (en) | Online road surface unevenness identification method and system | |
| CN112036598A (en) | Charging pile use information prediction method based on multi-information coupling | |
| CN116613745A (en) | PSO-ELM electric vehicle charging load prediction method based on variation modal decomposition | |
| CN116451322A (en) | Bayesian optimization-based LSTM deep learning network mechanical prediction method | |
| CN115794805A (en) | Medium-low voltage distribution network measurement data supplementing method | |
| CN116110219A (en) | Traffic accident prediction method | |
| CN113989066A (en) | Galvanic pile energy consumption analysis based on big data | |
| CN113947904A (en) | Multi-scale short-term traffic flow prediction method based on S-G filtering and deep belief network | |
| CN113569460A (en) | Real vehicle fuel cell system state multi-parameter prediction method and device | |
| Wang et al. | Research on parking space prediction based on long short-term memory | |
| CN117648631B (en) | Power battery health state estimation method for electric automobile group | |
| CN113255725B (en) | Automobile sensor attack detection and repair method based on two-stage LSTM | |
| CN112215520B (en) | Multi-target fusion passing method and device, computer equipment and storage medium |
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 |