CN112700047A - Oil quantity loss prediction method based on BP neural network - Google Patents
Oil quantity loss prediction method based on BP neural network Download PDFInfo
- Publication number
- CN112700047A CN112700047A CN202011637963.4A CN202011637963A CN112700047A CN 112700047 A CN112700047 A CN 112700047A CN 202011637963 A CN202011637963 A CN 202011637963A CN 112700047 A CN112700047 A CN 112700047A
- Authority
- CN
- China
- Prior art keywords
- neural network
- duration
- data
- fuel consumption
- vehicle driving
- 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
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 37
- 239000000446 fuel Substances 0.000 claims abstract description 43
- 230000006399 behavior Effects 0.000 claims abstract description 36
- 238000007418 data mining Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 24
- 230000006870 function Effects 0.000 claims description 15
- 238000003062 neural network model Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 9
- 210000002569 neuron Anatomy 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 claims description 6
- 230000001133 acceleration Effects 0.000 claims description 4
- 230000007935 neutral effect Effects 0.000 claims description 4
- 238000012360 testing method Methods 0.000 abstract description 6
- 238000012417 linear regression Methods 0.000 description 16
- 238000002474 experimental method Methods 0.000 description 5
- 238000012843 least square support vector machine Methods 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 239000002245 particle Substances 0.000 description 3
- 238000002790 cross-validation Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000005284 excitation Effects 0.000 description 2
- 230000003631 expected effect Effects 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000008450 motivation Effects 0.000 description 1
- 230000001953 sensory effect Effects 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
-
- G06Q50/40—
Abstract
The invention belongs to the technical field of data mining and automobile testing, and particularly relates to an oil mass loss prediction method based on a BP neural network. The method comprises the following steps: s1, acquiring vehicle driving behavior data through a vehicle sensor, and transmitting the data back to a main system through the vehicle sensor; s2, standardizing the vehicle driving behavior data to obtain a regression problem data set of the vehicle driving behavior data; s3, establishing an oil mass loss prediction model by using a regression method based on a BP neural network; and S4, obtaining fuel consumption prediction results based on different vehicle driving behaviors. The method has the characteristics of capability of effectively processing the fuel consumption prediction problem of the intelligent vehicle and high prediction precision.
Description
Technical Field
The invention belongs to the technical field of data mining and automobile testing, and particularly relates to an oil mass loss prediction method based on a BP neural network.
Background
In the process of automobile driving, an effective mathematical model is difficult to establish aiming at the complex relation between the driving behavior index of a driver and the oil consumption of a driving vehicle, so that in the process of vehicle oil consumption testing, the driving behaviors of different drivers are evaluated mainly through the subjective feeling of a back office. Obviously, the above evaluation mode completely depending on the manual supervisor experience greatly affects the scientificity of the evaluation result, and cannot meet the requirement of enterprises on rapid prediction of fuel consumption indexes of different driving behaviors.
In order to solve the problems of strong subjectivity and low efficiency in the evaluation process, scholars begin to use a data mining method to predict the oil consumption of different driving vehicles, strive to extract the mapping rules of vehicle sensor indexes and oil consumption indexes from a large amount of vehicle driving data, and assist or replace logistics personnel to complete sensory prediction of vehicle oil consumption. At present, the problem of intelligent fuel consumption prediction in vehicle driving is mainly solved by a linear regression method.
However, the vehicle driving history data is complex, and some difficulties are encountered by adopting linear regression, for example, sometimes in the regression analysis, what kind of factor is selected and what kind of expression is adopted for the factor is only a conjecture, which affects the diversity of the factor and the immeasurability of some factors, so that the regression analysis is limited in some cases, and the expected effect is often not obtained when dealing with the fuel consumption prediction problem. Therefore, it is necessary to design a method with better prediction accuracy and capable of effectively handling the problem of predicting the fuel consumption of the intelligent vehicle.
For example, in a fuel consumption prediction method based on a least square support vector machine described in chinese patent application No. CN201710453070.6, an improved particle swarm algorithm is used to optimize kernel function parameters and penalty factors of a least square support vector machine model, and a trained improved particle swarm algorithm is used to optimize a vehicle fuel consumption prediction model of the least square support vector machine to predict fuel consumption of a test sample. Although the fuel consumption sensitive characteristic parameters can be screened out by adopting the symmetric uncertainty, the accurate kernel function parameters and punishment factors can be obtained by adopting the improved particle swarm algorithm, the prediction precision of the least square support vector machine model is improved, and the defect that the deviation between the type authentication fuel consumption and the actual fuel consumption of the vehicle is large is effectively overcome, but the method has the defects that the application range is small and the prediction precision is low due to the fact that the least square support vector machine model is adopted, the method is difficult to help enterprises to improve the working efficiency in vehicle fuel consumption prediction, and helps the enterprises to scientifically and efficiently perform vehicle maintenance and driver excitation.
Disclosure of Invention
The invention provides an oil consumption prediction method based on a BP neural network, which can effectively solve the oil consumption prediction problem of an intelligent vehicle and has high prediction precision, and aims to overcome the problems that in the prior art, the expected effect cannot be obtained and the prediction precision is poor when the oil consumption of different driving vehicles is predicted by adopting a linear regression analysis mode.
In order to achieve the purpose, the invention adopts the following technical scheme:
the oil mass loss prediction method based on the BP neural network comprises the following steps:
s1, acquiring vehicle driving behavior data through a vehicle sensor, and transmitting the data back to a main system through the vehicle sensor;
s2, standardizing the vehicle driving behavior data to obtain a regression problem data set of the vehicle driving behavior data;
s3, establishing an oil mass loss prediction model by using a regression method based on a BP neural network;
and S4, obtaining fuel consumption prediction results based on different vehicle driving behaviors.
Preferably, the vehicle driving behavior data in step S2 includes an air conditioner operation duration, a heater operation duration, a green zone duration, a pulse amount, an idle speed duration, a driving duration, an engine operation duration, a GPS mileage, a pulse mileage, an excessively long idle speed number, an excessively long idle speed duration, a rapid acceleration number, a rapid acceleration duration, a rapid deceleration number, a rapid deceleration duration, an over-revolution number, an over-revolution duration, and an idle air conditioner number, the system comprises an idle air conditioner time length, a neutral gear sliding time length, a fatigue driving time length, a long time braking time length, a long time clutch time length, a violent accelerator stepping time length, a large accelerator stepping time length, a parking accelerator stepping time length, a cold vehicle driving time length, a parking immediate flameout time length, a long brake time length, a long clutch time length, a sudden accelerator stepping time length, a large accelerator stepping time length, a parking accelerator stepping time length, an immediate starting time length and an immediate parking time length.
Preferably, the process of establishing the fuel quantity loss prediction model in step S3 is as follows:
according to historical data of vehicle driving behaviors, a BP neural network model is established by using a data mining technology, and the generated BP neural network model is used for predicting the oil consumption of hundreds of kilometers of different vehicle driving behaviors.
Preferably, step S3 includes the steps of:
s31, collecting historical driving data of all commercial vehicles in a certain month, and establishing a training data sample set of the oil quantity loss prediction method;
s32, preprocessing a training data sample set of the oil mass loss prediction model, wherein the preprocessing is min-max standardization processing;
and S33, establishing an oil quantity loss prediction model by using a data mining technology of a feedback neural network (BPNN).
Preferably, the training data sample set of the fuel consumption prediction method comprises vehicle driving behavior data and operating vehicle fuel consumption per kilometer data.
Preferably, step S4 includes the steps of:
and (4) aiming at the unknown sample, according to the oil mass loss prediction model prediction result in the step S3 and outputting the result, and obtaining the fuel consumption per hundred kilometers prediction result of the vehicle based on different vehicle driving behaviors.
Preferably, step S33 further includes the steps of:
s331, neural network initialization, given training numberAccording to the data sample set (X, Y), the number M of nodes of an input layer, the number P of nodes of an implicit layer and the number N of nodes of an output layer are given, and a weight value v is initializedijAnd ωjkInitializing the hidden layer threshold αjOutput layer threshold betakGiving a learning rate t, giving a neuron activation function and giving an iteration number S;
the BP neural network adopts a single hidden layer structure, the number S of selected iteration is 100, the number P of hidden layer nodes is 10, the given learning rate t is 0.01, and the initialization weighted value and the threshold value are random numbers between 0 and 1; the neuron activation function is a Relu function;
s332, calculating hidden layer output H and output layer output O by the initialized weight value and the neuron activation function;
s333, calculating an error between the output layer O and the actual output Y of the training data sample, wherein the defined error E is as follows:
ek＝yk-ok k＝1,2,...,N；
s334, updating the weighted value v according to the error EijAnd ωjk：
ωjk＝ωjk+thjek j＝1,2,...,P；k＝1,2,...,N；
S335, updating the threshold value alpha according to the error E and the weight valuejAnd betak：
βk＝βk+ek k＝1,2,...,N；
S336, judging whether the iteration times are reached, and finishing the training process if the iteration times are reached; if not, continuing training until the iteration number reaches S.
Compared with the prior art, the invention has the beneficial effects that: (1) the method is used for predicting the fuel consumption of the vehicle for one hundred kilometers caused by different driving behaviors based on the BP neural network, and helping enterprises to establish an intelligent vehicle fuel consumption prediction system; (2) for the logistics personnel, the method can be utilized, the relevant sensor indexes of the vehicle are used as the input variables of the model, the model automatically outputs the predicted oil consumption, the prediction can be carried out with higher precision, and the logistics personnel can be helped to make better decisions in vehicle maintenance and driver motivation; (3) the BP neural network provided by the invention can effectively process the fuel consumption prediction problem of the intelligent vehicle, and compared with the traditional linear regression, the BP neural network can obtain better prediction precision; (4) the method can help enterprises to improve the working efficiency in vehicle oil consumption prediction and help the enterprises to scientifically and efficiently perform vehicle maintenance and driver excitation.
Drawings
Fig. 1 is a flowchart of a method for predicting fuel consumption based on a BP neural network according to the present invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention, the following description will explain the embodiments of the present invention with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
Example 1:
as shown in fig. 1, the present invention provides a method for predicting oil consumption based on a BP neural network, comprising the following steps:
step S1, collecting vehicle driving behavior data through a vehicle sensor, and transmitting the data back to a main system through the vehicle sensor;
the vehicle driving behavior data comprises air conditioner running time, heater running time, green zone time, pulse quantity, idle speed time, running time, engine running time, GPS mileage, pulse mileage, overlength idle times, overlength idle time, urgent acceleration times, urgent deceleration times, overtaking times, idle air conditioner times, neutral gear sliding times, fatigue driving times, long-time brake times, long-time clutch times, accelerator slamming times, large accelerator times, accelerator stopping times, cold vehicle running times, immediate flameout times, long brake times, long clutch times, accelerator slamming times, large accelerator stopping times, immediate accelerator starting time and immediate parking time. The above-mentioned 37 physical indexes in total are used as input variables, i.e., vehicle sensor indexes.
Step S2, standardizing the vehicle driving behavior data to obtain a regression problem data set of the vehicle driving behavior data;
step S3, establishing an oil mass loss prediction model by using a regression method based on a BP neural network;
the process of establishing the oil quantity loss prediction model is as follows:
according to historical data of vehicle driving behaviors, a BP neural network model is established by using a data mining technology, and the generated BP neural network model is used for predicting the oil consumption of hundreds of kilometers of different vehicle driving behaviors.
Step S3 specifically includes the following steps:
step S31, collecting historical driving data of all commercial vehicles in a month, and establishing a training data sample set of the oil quantity loss prediction method;
the training data sample set of the oil mass loss prediction method comprises vehicle driving behavior data and operating vehicle hundred kilometer oil consumption data;
the method comprises the steps of collecting historical driving data of a service vehicle from a certain enterprise to establish a fuel consumption prediction training data sample set for regression prediction, sorting the historical data, deleting some repeated or missing data, and finally obtaining each group of historical data including 37 physical indexes, wherein the fuel consumption prediction training data sample set of the embodiment includes service vehicle running data of the enterprise in 9 months in 2020 and 286 groups in total.
Step S32, preprocessing a training data sample set of the fuel quantity loss prediction model, wherein the preprocessing is min-max standardization processing;
definition set X ═ { X1,x2,...,x286Taking the value of the condition attribute, max (X) the maximum value in the set, and min (X) the minimum value in the set, then taking xnN ∈ {1, 2., 286} has a normalized result x'n：
And step S33, establishing an oil quantity loss prediction model by using a data mining technology of a feedback neural network (BPNN).
This example compares two different regression algorithms, including multiple Linear Regression (LR) and feedback neural network (BPNN).
The steps of modeling with a feedback neural network (BPNN) are as follows:
step S331, neural network initialization, given training data sample set (X, Y), given input layer node number M, hidden layer node number P, output layer node number N, initialization weight value vijAnd ωjkInitializing the hidden layer threshold αjOutput layer threshold betakGiving a learning rate t, giving a neuron activation function and giving an iteration number S;
the BP neural network adopts a single hidden layer structure, the number S of selected iteration is 100, the number P of hidden layer nodes is 10, the given learning rate t is 0.01, and the initialization weighted value and the threshold value are random numbers between 0 and 1; the neuron activation function is a Relu function;
step S332, calculating hidden layer output H and output layer output O by the initialized weight value and the neuron activation function;
step S333, calculating an error between the output layer output O and the training data sample actual output Y, and defining an error E as:
ek＝yk-ok k＝1,2,...,N；
step S334, updating the weighted value v according to the error EijAnd ωjk：
ωjk＝ωjk+thjek j＝1,2,...,P；k＝1,2,...,N；
Step S335, updating the threshold value alpha according to the error E and the weighted valuejAnd betak：
βk＝βk+ek k＝1,2,...,N；
Step S336, judging whether the iteration times are reached, and finishing the training process if the iteration times are reached; if not, continuing training until the iteration number reaches S.
The procedure for modeling using multiple Linear Regression (LR) is as follows:
step S337, establishing a multiple linear regression model:
f(xi)＝wTxi+b,
wherein xi＝(x1；x2；…；xd) Representing the vector of independent variables, wi＝(w1；w2；…；wd) Representing the multivariate Linear regression yields the parameters for the independent variable xiT denotes the transpose of the vector, bi＝(b1；b2；…；bd) Representing a constant vector.
In step S338, w and b are estimated by the least square method. For ease of discussion, w and b are absorbed into vector form
Accordingly, the data set D is represented as a matrix X of size n × (D +1), where each row corresponds to an example, the first D elements of the row correspond to the D attribute values of the example, and the last element is constantly set to 1. The mark is also written in vector form y ═ y (y)1；y2；…；yn) The following relationship exists:
where T represents the transpose of the vector.
Order to
To pair
Derived to obtain
Make the above formula zero available
Closed-form solution of the optimal solution. After w and b are found, the model is determined.
And S4, obtaining fuel consumption prediction results based on different vehicle driving behaviors.
And (4) aiming at the unknown sample, according to the oil mass loss prediction model prediction result in the step S3 and outputting the result, and obtaining the fuel consumption per hundred kilometers prediction result of the vehicle based on different vehicle driving behaviors.
In order to verify the performance of the present invention, in this embodiment, 286 sets of data samples provided by an enterprise are subjected to preprocessing and then are subjected to an experiment in a cross validation manner of three times and ten folds: during the experiment, all data samples are randomly divided into ten parts, nine parts of the data samples are taken as a training data sample set each time, the rest part of the data samples is taken as a data sample set to be evaluated, a regression method is adopted to calculate a classification result and compare the classification result with an actual result, and the accuracy, the average relative error, the maximum absolute value error and the average square error are obtained. This experiment was repeated three times and the results averaged ten times to obtain the final regression accuracy, average relative error, maximum absolute error and average squared error. The regression prediction evaluation results of the multivariate linear model and the three-fold ten-fold cross validation experiment based on the BP neural network method are shown in tables 1,2 and 3 below.
TABLE 1 Linear regression model and BP neural network model Fuel consumption prediction comparison
TABLE 2 comparison of Linear regression model and BP neural network model for fuel consumption prediction
TABLE 3 comparison of Linear regression model and BP neural network model for fuel consumption prediction
The specific calculation formula of the evaluation index in the table is as follows:
relative error (Ae):
wherein y refers to the actual fuel consumption value of the test sample,
the fuel consumption is predicted according to the fuel consumption value,
refers to the average value of the actual oil consumption.
Average relative error (Are):
maximum absolute value error (Maximum absolute error, Mae):
mean square error (Mean square error, Mse):
in tables 1,2 and 3, the bold part shows the optimum value of each column, so we can clearly see that the method adopted by the present invention is significantly superior to the conventional linear regression model in each test.
Compared with the traditional regression method, the method based on the invention has obvious advantages in vehicle oil consumption prediction, and can be used as an effective method for enterprise vehicle oil consumption prediction.
The BP neural network on which the present invention is based essentially implements a mapping function from input to output, whereas mathematical theories have demonstrated that the BP neural network has the capability of implementing any complex non-linear mapping, which makes it particularly suitable for solving problems with complex internal mechanisms. The method utilizes the BP neural network model to combine with actual vehicle data to establish a vehicle oil consumption prediction model.
According to the method, the vehicle oil consumption prediction is decomposed into the multilayer neural network, so that the problems that a multi-classification problem model is complex and difficult to solve can be effectively solved;
the invention designs a complete experiment to verify the effectiveness of the strategy.
The result of this embodiment shows that, compared with the classical linear regression prediction model, the prediction accuracy of the strategy of the present invention is significantly higher than that of the classical linear regression prediction model. By using the method, the fuel consumption prediction of the driving behavior of the logistics personnel can be assisted in the vehicle fuel consumption management process of an enterprise.
The foregoing has outlined rather broadly the preferred embodiments and principles of the present invention and it will be appreciated that those skilled in the art may devise variations of the present invention that are within the spirit and scope of the appended claims.
Claims (7)
1. The oil mass loss prediction method based on the BP neural network is characterized by comprising the following steps:
s1, acquiring vehicle driving behavior data through a vehicle sensor, and transmitting the data back to a main system through the vehicle sensor;
s2, standardizing the vehicle driving behavior data to obtain a regression problem data set of the vehicle driving behavior data;
s3, establishing an oil mass loss prediction model by using a regression method based on a BP neural network;
and S4, obtaining fuel consumption prediction results based on different vehicle driving behaviors.
2. The BP neural network-based fuel consumption prediction method according to claim 1, wherein the vehicle driving behavior data in step S2 includes an air conditioner operation duration, a heater operation duration, a green zone duration, a pulse quantity, an idle duration, a driving duration, an engine operation duration, a GPS mileage, a pulse mileage, an excessively long idle number, an excessively long idle duration, a rapid acceleration number, a rapid deceleration duration, an overrun number, an overrun duration, an idle air conditioner number, an idle air conditioner duration, a neutral coasting number, a neutral coasting duration, a fatigue driving duration, a long time brake number, a long time clutch number, a hard accelerator pedal number, a throttle pedal number, a cold vehicle driving number, a stop immediate shut-off number, a long brake duration, a long clutch duration, a fatigue driving duration, a long time brake number, a long time clutch number, a long time accelerator pedal number, a stop accelerator pedal number, The accelerator pedal pressing time length is a sudden accelerator pedal pressing time length, a large accelerator pedal pressing time length, a parking accelerator pedal pressing time length, an immediate starting time length and an immediate parking time length.
3. The method for predicting the fuel consumption based on the BP neural network according to claim 2, wherein the process of establishing the fuel consumption prediction model in the step S3 is as follows:
according to historical data of vehicle driving behaviors, a BP neural network model is established by using a data mining technology, and the generated BP neural network model is used for predicting the oil consumption of hundreds of kilometers of different vehicle driving behaviors.
4. The method for predicting the oil mass loss based on the BP neural network as claimed in claim 3, wherein the step S3 comprises the following steps:
s31, collecting historical driving data of all commercial vehicles in a certain month, and establishing a training data sample set of the oil quantity loss prediction method;
s32, preprocessing a training data sample set of the oil mass loss prediction model, wherein the preprocessing is min-max standardization processing;
and S33, establishing an oil quantity loss prediction model by using a data mining technology of a feedback neural network (BPNN).
5. The BP neural network-based fuel consumption prediction method according to claim 4, wherein the training data sample set of the fuel consumption prediction method comprises vehicle driving behavior data and operating vehicle one hundred kilometers fuel consumption data.
6. The method for predicting the oil mass loss based on the BP neural network as claimed in claim 5, wherein the step S4 comprises the following steps:
and (4) aiming at the unknown sample, according to the oil mass loss prediction model prediction result in the step S3 and outputting the result, and obtaining the fuel consumption per hundred kilometers prediction result of the vehicle based on different vehicle driving behaviors.
7. The method for predicting the oil quantity loss based on the BP neural network as claimed in any one of claims 4-6, wherein the step S33 further comprises the following steps:
s331, initializing the neural network, and giving training data samplesSet (X, Y), given number of nodes of input layer M, number of nodes of hidden layer P and number of nodes of output layer N, initializing weighted value vijAnd ωjkInitializing the hidden layer threshold αjOutput layer threshold betakGiving a learning rate t, giving a neuron activation function and giving an iteration number S;
the BP neural network adopts a single hidden layer structure, the number S of selected iteration is 100, the number P of hidden layer nodes is 10, the given learning rate t is 0.01, and the initialization weighted value and the threshold value are random numbers between 0 and 1; the neuron activation function is a Relu function;
s332, calculating hidden layer output H and output layer output O by the initialized weight value and the neuron activation function;
s333, calculating an error between the output layer O and the actual output Y of the training data sample, wherein the defined error E is as follows:
ek＝yk-ok k＝1,2,...,N；
s334, updating the weighted value v according to the error EijAnd ωjk：
ωjk＝ωjk+thjek j＝1,2,...,P；k＝1,2,...,N；
S335, updating the threshold value alpha according to the error E and the weight valuejAnd betak：
βk＝βk+ek k＝1,2,...,N；
S336, judging whether the iteration times are reached, and finishing the training process if the iteration times are reached; if not, continuing training until the iteration number reaches S.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011637963.4A CN112700047A (en) | 2020-12-31 | 2020-12-31 | Oil quantity loss prediction method based on BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011637963.4A CN112700047A (en) | 2020-12-31 | 2020-12-31 | Oil quantity loss prediction method based on BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112700047A true CN112700047A (en) | 2021-04-23 |
Family
ID=75513938
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011637963.4A Pending CN112700047A (en) | 2020-12-31 | 2020-12-31 | Oil quantity loss prediction method based on BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112700047A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113642227A (en) * | 2021-06-29 | 2021-11-12 | 桂林电子科技大学 | BP neural network oil consumption prediction method based on fusion driving behavior characteristics |
| CN114608604A (en) * | 2022-05-12 | 2022-06-10 | 山东捷瑞数字科技股份有限公司 | Vehicle oil consumption prediction method and device based on machine learning and GPS positioning |
| CN117235508A (en) * | 2023-11-15 | 2023-12-15 | 天津市普迅电力信息技术有限公司 | Vehicle fuel consumption prediction method and system based on big data |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104200267A (en) * | 2014-09-23 | 2014-12-10 | 清华大学 | Vehicle driving economy evaluation system and vehicle driving economy evaluation method |
| CN108711016A (en) * | 2018-06-05 | 2018-10-26 | 合肥湛达智能科技有限公司 | A kind of driving behavior methods of marking based on BP neural network |
-
2020
- 2020-12-31 CN CN202011637963.4A patent/CN112700047A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104200267A (en) * | 2014-09-23 | 2014-12-10 | 清华大学 | Vehicle driving economy evaluation system and vehicle driving economy evaluation method |
| CN108711016A (en) * | 2018-06-05 | 2018-10-26 | 合肥湛达智能科技有限公司 | A kind of driving behavior methods of marking based on BP neural network |
Non-Patent Citations (1)
| Title |
|---|
| 高涛 等: "基于神经网络的驾驶行为与油耗相关性分析", 计算机与数字工程, no. 05, pages 803 - 807 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113642227A (en) * | 2021-06-29 | 2021-11-12 | 桂林电子科技大学 | BP neural network oil consumption prediction method based on fusion driving behavior characteristics |
| CN114608604A (en) * | 2022-05-12 | 2022-06-10 | 山东捷瑞数字科技股份有限公司 | Vehicle oil consumption prediction method and device based on machine learning and GPS positioning |
| CN117235508A (en) * | 2023-11-15 | 2023-12-15 | 天津市普迅电力信息技术有限公司 | Vehicle fuel consumption prediction method and system based on big data |
| CN117235508B (en) * | 2023-11-15 | 2024-01-30 | 天津市普迅电力信息技术有限公司 | Vehicle fuel consumption prediction method and system based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112700047A (en) | Oil quantity loss prediction method based on BP neural network | |
| CN107688343B (en) | Energy control method of hybrid power vehicle | |
| CN109635830B (en) | Screening method for effective data for estimating automobile quality | |
| Wong et al. | Automotive engine power performance tuning under numerical and nominal data | |
| CN112026744B (en) | Series-parallel hybrid power system energy management method based on DQN variants | |
| DE102020106591A1 (en) | Method for creating a trained artificial neural network, method for predicting emission data of a vehicle and method for determining calibration values | |
| CN111456840A (en) | Intelligent control method for cooling water flow of internal combustion engine based on RBF neural network | |
| CN113988705A (en) | Traffic safety risk assessment method and device | |
| DE102020215350A1 (en) | MODELING OPERATIONAL PROFILES OF A VEHICLE | |
| Wei | Design of Web based expert system of electronic control engine fault diagnosis | |
| CN109238740B (en) | Comprehensive evaluation method for automobile sliding resistance based on whole automobile mass and volume | |
| Kolachalama et al. | Using Deep Learning to Predict the Engine Operating Point in Real-Time | |
| CN116741304A (en) | Diesel engine original exhaust nitrogen oxide prediction method based on Internet of vehicles big data | |
| Kan et al. | A deep learning engine power model for estimating the fuel consumption of heavy-duty trucks | |
| Kim et al. | Application of neural network model to vehicle emissions | |
| Zhou et al. | Objective evaluation of drivability in passenger cars with dual-clutch transmission: a case study of static gearshift condition | |
| Wang et al. | A novel fuzzy comprehensive evaluation model for application effect of connected vehicle system in a tunnel scenario | |
| Witaszek | Modeling of fuel consumption using artificial neural networks | |
| DE102015102951A1 (en) | Method for controlling a cooling system in a vehicle | |
| CN114943403A (en) | Bus driver ecological driving evaluation system and method based on big data | |
| Weber et al. | Data-driven bev modeling for realistic consumption calculation in traffic simulation | |
| Liu et al. | A method to improve accuracy of velocity prediction using Markov model | |
| Zhang | Vehicle Parameters Estimation and Driver Behavior Classification for Adaptive Shift Strategy of Heavy Duty Vehicles | |
| Peng et al. | Construction of engine emission test driving cycle of city transit buses | |
| Bhattacharjee et al. | Estimation of driver demand torque using parametric and nonparametric data-driven 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 |