CN112668799A - Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data - Google Patents
Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data Download PDFInfo
- Publication number
- CN112668799A CN112668799A CN202110001963.3A CN202110001963A CN112668799A CN 112668799 A CN112668799 A CN 112668799A CN 202110001963 A CN202110001963 A CN 202110001963A CN 112668799 A CN112668799 A CN 112668799A
- Authority
- CN
- China
- Prior art keywords
- vehicle speed
- vehicle
- prediction
- driving
- time
- 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
- 238000007726 management method Methods 0.000 title claims abstract description 36
- 238000003860 storage Methods 0.000 title claims abstract description 11
- 238000005457 optimization Methods 0.000 claims abstract description 50
- 238000000034 method Methods 0.000 claims abstract description 24
- 230000007774 longterm Effects 0.000 claims abstract description 21
- 238000005265 energy consumption Methods 0.000 claims abstract description 19
- 230000008569 process Effects 0.000 claims abstract description 15
- 241000156302 Porcine hemagglutinating encephalomyelitis virus Species 0.000 claims abstract 3
- 230000001133 acceleration Effects 0.000 claims description 21
- 238000013528 artificial neural network Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 8
- 238000012937 correction Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000005096 rolling process Methods 0.000 claims description 4
- 230000014759 maintenance of location Effects 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 description 13
- 239000000446 fuel Substances 0.000 description 12
- 238000005070 sampling Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 210000002569 neuron Anatomy 0.000 description 5
- 238000009826 distribution Methods 0.000 description 4
- 238000003064 k means clustering Methods 0.000 description 4
- 230000007704 transition Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003638 chemical reducing agent Substances 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Landscapes
- Electric Propulsion And Braking For Vehicles (AREA)
Abstract
The invention discloses an intelligent energy management method and a storage medium of a PHEV (Power vehicle Environment) based on big driving data, wherein the method comprises the steps of obtaining historical driving data of a vehicle to obtain a vehicle speed characteristic parameter and a road characteristic parameter; establishing a long-term vehicle speed prediction model; acquiring real-time road characteristic parameters, predicting long-time vehicle speed to obtain a first prediction result, and predicting driving energy consumption by using the first prediction result; establishing a short-time vehicle speed prediction model; acquiring real-time vehicle speed characteristic parameters, carrying out short-time vehicle speed prediction to obtain a second prediction result, carrying out model prediction control by using the second prediction result, and carrying out instantaneous optimization on the power of an engine and a motor; and updating the global energy consumption plan at fixed time intervals in the driving process. The invention creatively provides a multi-level energy management strategy combining plug-in hybrid electric vehicle energy global planning and power instantaneous optimization control, so as to reduce the energy cost in the PHEV driving process.
Description
Technical Field
The invention relates to the field of intelligent energy management of intelligent networked automobile individuals, in particular to an intelligent energy management method and a storage medium of a PHEV (Plug-in hybrid electric vehicle) based on big driving data.
Background
The energy management strategy improves the fuel economy and emission performance of the whole vehicle by coordinating the distribution mode of fuel energy and electric energy of the hybrid electric vehicle, and particularly relates to the optimization problem of the output power or torque of an engine and a motor in a full path. Energy management policies are largely divided into rule-based policies and optimization-based policies, depending on the control implementation. The dynamic characteristics of the management strategy based on the rules are poor, and the matching effect of the power system is not high. The strategies based on optimization are divided into management strategies based on global optimization, management strategies based on instantaneous optimization and management strategies based on local optimization, wherein the global optimization can reach a global theoretical optimal value generally, but the practicability is low because the global condition cannot be obtained and the calculated amount is large; the instantaneous optimization seeks optimal distribution on the power or torque of the engine and the motor aiming at the minimum energy consumption at the current moment, the future running information of the vehicle does not need to be known, the calculated amount is small, and the overall optimal control effect cannot be guaranteed; the global optimization problem is decomposed into segmented local optimization problems by predicting the future short-time driving condition based on the strategy of local optimization and applying a model prediction control method, so that the feasibility of the global optimization strategy is improved.
Disclosure of Invention
In order to solve the deficiency of global optimization, instantaneous optimization is added in global planning, the power distribution of a PHEV engine and a motor is rapidly coordinated, and the purpose of improving the fuel economy is achieved. Therefore, the invention provides an intelligent energy management method and a storage medium of a PHEV (Power vehicle) based on big driving data, which realize a multilevel energy management strategy combining the energy global planning and the power instantaneous optimization of the PHEV so as to improve the fuel economy of a vehicle-mounted power system during the driving of the PHEV.
The intelligent energy management method of the PHEV based on the big driving data comprises the following steps,
obtaining historical driving data of a vehicle, and processing the historical driving data to obtain a vehicle speed characteristic parameter and a road characteristic parameter;
establishing a long-term vehicle speed prediction model; acquiring real-time road characteristic parameters, predicting long-time vehicle speed to obtain a first prediction result, predicting driving energy consumption by using the first prediction result, outputting predicted power, driving energy consumption and available electric energy of a power battery, and determining the working mode of a power system of the plug-in hybrid electric vehicle;
establishing a short-time vehicle speed prediction model; acquiring real-time vehicle speed characteristic parameters, carrying out short-time vehicle speed prediction to obtain a second prediction result, carrying out model prediction control by using the second prediction result, and carrying out instantaneous optimization on the power of an engine and a motor;
and updating the global energy consumption plan at fixed time intervals in the driving process.
According to the technical scheme, the model is used for predicting and controlling instantaneous optimization fuel consumption and the battery charge state, and energy point optimization is converted into interval optimization, so that a multi-level energy management strategy combining energy global planning and power instantaneous optimization of the PHEV is realized, and the energy cost of the PHEV is reduced.
Further, the obtaining of the historical vehicle driving data and the processing of the historical vehicle driving data to obtain the vehicle speed characteristic parameters and the road characteristic parameters are specifically that the historical vehicle driving data including position information and vehicle driving information are obtained by using a vehicle-mounted positioning system and a vehicle driving state sensor, and the vehicle driving information includes a traffic state, a vehicle speed and a vehicle distance; the vehicle speed characteristic parameters comprise an acceleration proportion and a deceleration proportion, and the road characteristic parameters comprise single driving mileage, road type, road speed limit, traffic congestion degree, average vehicle distance and average traffic light distance.
Further, the establishing of the long-term vehicle speed prediction model specifically includes that an LSTM neural network is adopted to establish the long-term vehicle speed prediction model, the long-term vehicle speed prediction model includes an input gate, a forgetting gate and an output gate, the input gate controls the input of the road characteristic parameters, the forgetting gate controls the retention of the historical state information, and the output gate is used for outputting a long-term vehicle speed prediction sequence.
The LSTM (Long Short-Term Memory network) is a time recurrent neural network, can overcome the problem that the traditional recurrent network excessively depends on a Long-Term sequence, and has strong learning ability on the time sequence.
Furthermore, the output of the driving energy consumption prediction by utilizing the first prediction result also comprises an acceleration characteristic ratio correction term,wherein v is vehicle speed; dv/dt is the linear acceleration; t is the acceleration and deceleration characteristic proportion.
Further, the available electric energy E of the power batterymF (D, SOC, T), where D is road condition, SOC is battery state of charge, and T is battery temperature.
Further, the establishing of the short-time vehicle speed prediction model specifically includes dividing the driving condition types into a steady condition and a violent condition, predicting the vehicle speed by using a Markov chain model under the steady condition, and predicting the vehicle speed by using a BP neural network model under the violent condition.
Under the stable working condition, the acceleration of the vehicle at each moment is approximately independent of historical data, so that a Markov chain model can be used for simulating the speed change rule and predicting the future speed; under the violent working conditions, the vehicle speed characteristic parameters are changed violently and fluctuate greatly, so that the BP neural network is adopted to learn the behavior of the driver to predict the future vehicle speed.
Further, the working modes of the hybrid electric vehicle power system comprise an electric power mode, a pure engine mode, a hybrid driving mode and a driving charging mode.
Further, the instantaneous optimization of the power of the engine and the motor is specifically that the energy global optimal control under the whole driving working condition is converted into local optimal control in a prediction region, the running state or control parameters in the next time domain are updated through continuous rolling optimization so as to obtain an optimization result, and the high robustness of the global optimization and the real-time performance of the instantaneous optimization are combined to achieve the optimal management of the energy of the whole vehicle.
The present invention also provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the method according to any of the above-mentioned embodiments when running.
The invention has the beneficial effects that: according to the invention, the LSTM neural network is adopted to predict the long-term vehicle speed, and a driving energy consumption model is established based on the long-term vehicle speed prediction sequence, so that the energy global planning is realized. The working conditions are divided into stable working conditions and violent working conditions, the working conditions of the vehicle in the running process are judged on line, the short-time vehicle speed prediction is carried out, and different prediction models are adopted for the two working conditions to improve the prediction accuracy. And performing instantaneous optimization on energy consumption by using model predictive control, and updating the global energy plan at fixed time intervals. The invention creatively provides a multi-level energy management strategy combining plug-in hybrid electric vehicle energy global planning and power instantaneous optimization control, so as to reduce the energy cost in the PHEV driving process.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a system framework diagram of an intelligent energy management method of a PHEV based on big driving data according to the invention;
FIG. 2 is a flowchart of an embodiment of an intelligent energy management method for a PHEV based on big driving data;
FIG. 3 is a diagram of an LSTM neural network architecture in an embodiment of the present invention;
FIG. 4 is a diagram of a BP neural network structure according to an embodiment of the present invention;
FIG. 5 is a flow chart of an embodiment of an energy management MPC transient optimization method.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In order to improve fuel economy and promote the development of electric automobiles, the invention provides a multi-level energy management strategy method combining plug-in hybrid electric automobile energy global planning and power instantaneous optimization control. Establishing a long-term vehicle speed prediction model by using an LSTM neural network based on the historical driving big data of the vehicle; obtaining road characteristic parameters on line to predict the long-term speed, and realizing energy global planning based on the predicted speed; dividing the working condition into a stable working condition and a violent working condition by using a K-means clustering algorithm; judging the working condition of the vehicle on line and predicting the future short-time vehicle speed by using different prediction models; the model is used for predicting and controlling instantaneous optimization fuel consumption and the battery charge state, and energy point optimization is converted into interval optimization, so that a multi-level energy management strategy combining energy global planning and power instantaneous optimization of the PHEV is realized, and the energy cost of the PHEV is reduced.
Referring to fig. 1, the technical solution of the present invention is an intelligent energy management method for a PHEV based on big driving data, comprising the following steps,
and S1, obtaining historical driving data of the vehicle, and processing the historical driving data to obtain vehicle speed characteristic parameters and road characteristic parameters.
Specifically, under the condition of no prior working condition information, the accurate prediction of the future vehicle speed by using the historical driving data and the current time data of the vehicle is the primary step of the energy management strategy. The embodiment utilizes the vehicle-mounted positioning system and the vehicle running state sensor to acquire vehicle historical running data, including position information and vehicle running information, wherein the vehicle running information comprises traffic state, vehicle speed and vehicle distance.
The embodiment selects the vehicle speed characteristic parameter as the single-time running plus deceleration running proportion T. The road characteristic parameters can be obtained from a vehicle-mounted positioning system, and comprise single-travel mileage, road type, road speed limit, traffic congestion degree, average vehicle distance and average traffic light distance.
S2, establishing a long-term vehicle speed prediction model;
particularly, the LSTM is a time recursive neural network, can overcome the problem that the traditional recursive network excessively depends on a long-time sequence, and has strong learning ability on the time sequence. The structure of the LSTM algorithm is shown in fig. 3.
The LSTM vehicle speed prediction algorithm structure consists of an input gate, a forgetting gate and an output gate. Wherein, the input door control information input comprises the road characteristic parameters mentioned in S1, namely single driving distance dis, road type road, road speed limit spl, traffic congestion degree tra, vehicle average distance inv and traffic light average distance ldis; the forgetting gate controls the retention of the historical state information, the output of the historical state information is a number in a [0,1] interval, 0 represents completely-discarded information, and 1 represents completely-retained information; the output gate is used to control the output of information, i.e., the long-term vehicle speed prediction sequence. The transfer relationship between the layers is as follows:
ht=Ot tanh(Ct)
wherein W and b _ l are parameter matrixes;inputting for the time t; h ist-1Output for the time t-1; h istOutput for time t; i.e. itThe input of an input gate at the moment t; f. oftThe output of the forgetting gate at the moment t; ctCell unit status at time t; o istThe output of the output gate at time t.
The independent variable t is difficult to determine when the long-time vehicle speed prediction model is acquired, so that the independent variable t needs to be converted into a fixed distance because the independent variable is a determined value in a regression prediction algorithm, otherwise, the dependent variable cannot be predicted, and the v-t long-time vehicle speed prediction model is converted into a v-s long-time vehicle speed prediction model by adopting an interpolation function interper1 carried by MATLAB. The transformation formula is as follows:
where dis is the single trip mileage and the spacing is 1m, i.e. dist+1-dist=1。
S3, acquiring real-time road characteristic parameters, predicting long-time vehicle speed to obtain a first prediction result, predicting driving energy consumption by using the first prediction result, outputting predicted power, driving energy consumption and available electric energy of a power battery, and determining the working mode of a power system of the plug-in hybrid electric vehicle;
specifically, the vehicle running power is calculated according to the long-time vehicle speed prediction sequence to obtain a predicted power value, and the vehicle running energy consumption can be obtained by integrating the predicted power value, wherein the calculation formula is as follows:
wherein, PdTo a predicted power; m is the loading quality of the whole vehicle; u. ofaAs the speed of a vehicle;CdIs the air resistance coefficient; a is the windward area; f is a rolling resistance coefficient; delta is a rotational inertia coefficient; du/dt is the linear acceleration.
In a preferred embodiment, the characteristic of predicting the long-term vehicle speed is that the accuracy of predicting the dynamic characteristics of the vehicle speed is not high, and the accuracy of predicting the energy is affected, so that the acceleration characteristic ratio correction term can be added at the same time of predicting and calculating the energy of the long-term vehicle speed. When acceleration is appliedThe vehicle belongs to an acceleration or deceleration running state. The calculation formula of the acceleration characteristic proportion correction term is as follows:
wherein, PTIs an acceleration proportional power correction term; v is vehicle speed; dv/dt is the linear acceleration; t is the acceleration and deceleration characteristic proportion.
The formula for calculating the running energy consumption of the whole vehicle is as follows:
wherein E is the energy consumption of the travelling crane.
The formula for calculating the available electric energy of the vehicle-mounted power battery is as follows:
Em=f(D,SOC,T)
wherein D is road working condition; SOC is the state of charge of the battery; and T is the battery temperature.
Road characteristic parameters are obtained in real time through a vehicle-mounted positioning system, long-term speed of remaining mileage is predicted on line, global energy planning is carried out on the PHEV through predicted power, and the working modes of an engine and a motor in the vehicle driving process are determined according to the following rules:
Em>E,0<Pr<Pelrunning in a pure electric mode;
Em<E,Pel<Pr<Pehpure engine mode travel;
Em<E,Pe_max<Prhybrid drive mode travel;
SOC<SOCmin&Peh<Pr<Pe_maxand driving in a driving charging mode.
Wherein, PrPower is required for the whole vehicle; pelMinimum power for the engine high efficiency zone; pehThe maximum power of the high-efficiency area of the engine; pe_maxThe maximum power of the engine; SOCminAnd the minimum SOC value of the power battery is obtained.
S4, establishing a short-time vehicle speed prediction model;
when the driving condition of the vehicle is divided, the commonly used algorithms include a K-means clustering algorithm, Expectation Maximization (EM) clustering of a Gaussian Mixture Model (GMM), a fuzzy clustering algorithm, and the like.
The K-means clustering algorithm classifies data by calculating the degree of affinity and sparseness among samples, and finally realizes that the data in the same class have larger characteristic similarity and the data in different classes have obvious difference.
Specifically, a plurality of cyclic working conditions are combined to form a sample, the vehicle speed characteristic parameter of the past 10s is calculated at each sampling moment in the cyclic working conditions, and a sample data set X is obtained1=[x11,x12,…,x1m],X2=[x21,x22,…,x2m],…,Xn=[xn1,xn2,…,xnm]Wherein X ═ X1,X2,…,XnAnd m is the number of characteristic parameters, and n is the length of the cycle condition.
Using a K-means clustering algorithm to randomly select a clustering center C1=[c11,c12,…,c1m],C2=[c21,c22,…,c2m]Wherein C ═ { C1,C2,…,CkAnd the Euclidean distance between the sample data and the clustering center:
wherein XiRepresenting the ith sample, i is more than or equal to 1 and less than or equal to n; cjJ is more than or equal to 1 and less than or equal to k and represents the jth clustering center; xitRepresenting the t dimension of the ith sample, wherein t is more than or equal to 1 and less than or equal to m; cjtRepresenting the t-dimension of the jth cluster center. Sequentially comparing the distance from each sample to each cluster center, and distributing the samples to the cluster of the cluster center closest to the sample center to form a new cluster S ═ S1,S2,…,Sk}. Then updating the clustering center according to the following formula:
wherein C islRepresenting the clustering center of the first cluster, wherein l is more than or equal to 1 and less than or equal to k; i SlL represents the number of samples in the ith class cluster; xiRepresents the ith sample in the ith cluster, 1 is more than or equal to i is less than or equal to | Sl|。
Iterating the steps, and when the change of the clustering center is smaller than a threshold value, determining that the classification is stable, and finally respectively obtaining the clustering centers C under stable working conditions1And clustering center C of violent working conditions2。
In the actual running process of the vehicle, the past 10s working condition characteristic parameter value [ x ] is calculated at the current sampling moment1,x2,…,xm]Calculating the characteristic parameter values to two cluster centers C1And C2Distance dist of1And dist2If:
dist1≤dist2judging that the current moment is a stable working condition;
dist1>dist2judging the current working condition to be violent working condition
And S5, acquiring real-time vehicle speed characteristic parameters, and performing short-time vehicle speed prediction to obtain a second prediction result.
Specifically, under a stable working condition, the acceleration of the vehicle at each moment is approximately independent of historical data, namely a Markov chain model can be used for simulating the speed change rule of the vehicle, and the future speed of the vehicle is predicted.
The Markov chain models are all composed of vehicle speed v (0-30m/s) and acceleration a (-1.5-1.5 m/s)2) Forming a discrete space. Dividing the current state quantity v into p spaces, and dividing the output quantity a at the next moment into q spaces to obtain a transition probability matrix of the Markov chain:
Tij=Pr[ak+n+1=aj|vk+n=vi]
wherein N is equal to {1, …, N ∈ [ ]pIs the predicted time-domain target time, TijFor the current moment of the vehicle speed vk+n=viThe acceleration of the vehicle satisfies a at the next momentk+n+1=ajThe transition probability of (2).
In the initial state, a Markov chain transition probability matrix is calculated according to the following formula:
based on the Markov chain model, the acceleration at the next moment can be predicted according to the vehicle speed at the current moment k, and the vehicle speed at the next moment is calculated:
by analogy, predicting the time domain NpThe vehicle speed at each moment can be calculated as follows:
wherein n is less than or equal to P, so that a vehicle speed prediction sequence under a stable working condition can be obtained.
Under the violent working conditions, the vehicle speed characteristic parameters are changed violently and fluctuate greatly, so that the BP neural network is adopted to learn the behavior of the driver to predict the future vehicle speed.
The BP neural network is a multi-level feedforward neural network, has the characteristics of information forward propagation and error backward propagation, the learning rule is a gradient descent method, in the learning process of the neural network, firstly, a sample carries out the forward propagation of information, the layer-by-layer processing is carried out according to the path sequence of an input layer, a hidden layer and an output layer, the error between the predicted output and the target output is obtained, the backward propagation stage is carried out, the weight value and the threshold value of the neural network are adjusted according to the predicted error, the predicted output of the neural network gradually approaches to the target output, and the BP neural network structure is shown in figure 4. The input X of the neural network is defined here as the driver pedal information and the historical vehicle speed 10s before the sampling time:
X=α1,α2,vk,vk-1,…,vk-9
wherein alpha is1Is accelerator pedal information; alpha is alpha2Is brake pedal information.
The model output is the predicted vehicle speed for the future 5 s:
Y=vk+1,vk+2,…,vk+5
firstly, initializing a network, and determining the number n of nodes of an input layer, an input vector X, the number q of nodes of a hidden layer, the number m of nodes of an output layer and an output vector Y of the network.
Hidden layer neuron input vjAnd an output hjThe calculation formula of (a) is as follows:
hj=f(vj)
wherein wijJ is more than or equal to 1 and less than or equal to q; a isjThe neuron threshold value of the hidden layer is set; f () is the hidden layer activation function.
Output layer neuron outputkThe calculation formula is as follows:
wherein, wjkConnecting weight between the hidden layer and the output layer; bkIs the neuron threshold of the output layer.
Updating the weight and the threshold of each layer in the BP algorithm error back propagation stage, wherein the updating formula is as follows:
wherein η is the learning rate; dkK is more than or equal to 1 and less than or equal to m, and is an error value between an actual value and a network output value.
The input quantity n of the BP neural network is 12, the output quantity m is 5, and the number q of the hidden layer neurons is 10. And in the running process of the vehicle, obtaining the characteristic parameter information of the vehicle in real time, and predicting the short-time vehicle speed on line.
And S6, performing model prediction control by using the second prediction result, and performing instantaneous optimization on the power of the engine and the motor.
MPC (model predictive control) is an open-loop optimal control strategy that solves a finite time domain at each sampling time, and is a control method that solves an optimal control sequence only at the next time by using the current state of the process state as the initial state of the optimal control problem. The model prediction control is divided into three steps of prediction model, rolling optimization and feedback correction, and a good real-time control effect can be achieved.
The embodiment carries out online real-time optimization of power based on a model predictive control algorithm, selects the engine torque as a system control quantity u, and adopts the engine torque TeAs a control amount; defining the system state quantity as x, and adopting the power battery SOC as the state quantity in the embodiment; and (3) setting a control system model with the system output quantity of y:
wherein the state quantity and the control quantity are respectively:
the relationship between the engine and motor torques and the total vehicle required torque is as follows:
Tr(k)=Te(k)+Tm(k)
wherein, Tr(k) The torque required by the whole vehicle at the moment k is obtained; t ise(k) Engine torque at time k; t ism(k) Motor torque at time k.
The whole vehicle required torque calculation formula is as follows:
wherein igIs the transmission ratio of the gearbox; i.e. i0The transmission ratio of the main speed reducer is set; etatThe comprehensive efficiency of the transmission system is obtained; and R is the radius of the wheel.
The relationship between the engine speed and the wheel speed is as follows:
ne(k)=nw(k)igi0
wherein n ise(k) The engine speed at the moment k; n isw(k) The wheel speed at time k.
The SOC state transition equation of the power battery is as follows:
wherein, UbatThe open circuit voltage of the power battery is obtained; pbatPower for the power battery; rbatThe internal resistance of the power battery; qbatIs the total capacity of the power battery.
In the embodiment, the fuel consumption is taken as the output quantity of the system, the minimum fuel consumption is taken as the optimization target, and the index function is as follows:
wherein f isk(SOC (k)) is an index function and is the minimum value of fuel consumption in the whole control process; n is the number of stages after the discrete prediction time domain; vk(SOC(k),Te(k) Is the fuel consumption of the k stages.
The constraint conditions of relevant parameters of the SOC of the engine, the motor and the power battery are as follows:
wherein n ism(k) The motor speed at time k.
When solving the optimization problem, firstly discretizing the system model, and adopting a dynamic programming algorithm to solve the solution of the optimization problem on line in real time, wherein the sequence of the optimal control quantity in the prediction time domain is as follows:
U*(k)=[u*(k),…,u*(k+P-1)]
the system control quantity at the current moment is as follows:
u(x(k))=u*(k)
the model predictive control is used in a PHEV energy management strategy, and vehicle information such as vehicle speed, battery SOC, engine rotating speed torque, motor rotating speed torque and the like is combined to optimize power distribution in real time, so that the fuel economy of the PHEV is improved. At each sampling time k, the model predictive control flowchart is shown in fig. 5, and the specific steps are as follows:
1) setting the prediction time length as p, and the prediction time domain at the moment k as [ k, k + p ];
2) the method comprises the steps of acquiring vehicle speed characteristic parameters in real time through a sensor, predicting the vehicle speed in a prediction time domain, and calculating the whole vehicle power and the required torque in the prediction time domain, wherein the power and the required torque can be expressed as [ T [ [ T ]r(k),Tr(k+p)];
3) Constructing a model prediction control optimization problem, and adopting a dynamic programming algorithm to solve an index function f in a prediction time domain on linek+p(SOC (k + p)) optimal control sequence;
4) applying a first set of optimal control quantities to the system model;
5) this process continues to be repeated at the next sampling instant;
and S7, updating the global energy consumption plan at fixed time intervals in the driving process.
In this embodiment, the energy global planning is updated every 5min in the driving process, and the MPC power instantaneous optimization is accompanied, so that a multi-level energy management strategy of the PHEV, which combines the power instantaneous optimization and the energy global planning based on the vehicle driving data, is realized. The above-mentioned time interval of 5min is only for illustration, and the time interval can be specifically determined according to the conditions of the vehicle and the road.
The working modes of the hybrid electric vehicle power system mentioned in the technical scheme comprise a pure electric mode, a pure engine mode, a hybrid driving mode and a running charging mode.
In another aspect, an embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, where the computer program is configured to, when executed, perform the steps in any of the above method embodiments.
Alternatively, in an embodiment of the present invention, the computer-readable storage medium may be configured to store a computer program for executing the above-described embodiment.
Optionally, in an embodiment of the present invention, the computer-readable storage medium may include, but is not limited to: various media capable of storing computer programs, such as a usb disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (9)
1. An intelligent energy management method of PHEV based on big driving data is characterized by comprising the following steps,
obtaining historical driving data of a vehicle, and processing the historical driving data to obtain a vehicle speed characteristic parameter and a road characteristic parameter;
establishing a long-term vehicle speed prediction model; acquiring real-time road characteristic parameters, predicting long-time vehicle speed to obtain a first prediction result, predicting driving energy consumption by using the first prediction result, outputting predicted power, driving energy consumption and available electric energy of a power battery, and determining the working mode of a power system of the plug-in hybrid electric vehicle;
establishing a short-time vehicle speed prediction model; acquiring real-time vehicle speed characteristic parameters, carrying out short-time vehicle speed prediction to obtain a second prediction result, carrying out model prediction control by using the second prediction result, and carrying out instantaneous optimization on the power of an engine and a motor; and updating the global energy consumption plan at fixed time intervals in the driving process.
2. The intelligent energy management method according to claim 1, wherein the obtaining of the historical vehicle driving data and the processing of the vehicle speed characteristic parameter and the road characteristic parameter are specifically implemented by obtaining the historical vehicle driving data, including position information and vehicle driving information, by using a vehicle-mounted positioning system and a vehicle driving state sensor, wherein the vehicle driving information includes a traffic state, a vehicle speed and a vehicle distance; the vehicle speed characteristic parameters comprise an acceleration proportion and a deceleration proportion, and the road characteristic parameters comprise single driving mileage, road type, road speed limit, traffic congestion degree, average vehicle distance and average traffic light distance.
3. The intelligent energy management method according to claim 2, wherein the establishing of the long-term vehicle speed prediction model is specifically that an LSTM neural network is adopted to establish the long-term vehicle speed prediction model, the long-term vehicle speed prediction model comprises an input gate, a forgetting gate and an output gate, the input gate controls the input of the road characteristic parameters, the forgetting gate controls the retention of the historical state information, and the output gate is used for outputting a long-term vehicle speed prediction sequence.
4. The intelligent energy management method according to claim 1, wherein the output of the driving energy consumption prediction using the first prediction result further comprises an acceleration characteristic ratio correction term:wherein v is vehicle speed; dv/dt is the linear acceleration; t is the acceleration and deceleration characteristic proportion.
5. The intelligent energy management method according to claim 1, wherein the electric energy E available for the power batterymF (D, SOC, T), where D is road condition, SOC is battery state of charge, and T is battery temperature.
6. The intelligent energy management method according to claim 1, wherein the establishing of the short-time vehicle speed prediction model is specifically to divide the types of driving conditions into a steady condition and a violent condition, predict the vehicle speed by using a Markov chain model under the steady condition, and predict the vehicle speed by using a BP neural network model under the violent condition.
7. The intelligent energy management method of claim 1, wherein the hybrid vehicle powertrain operating modes include an electric-only mode, an engine-only mode, a hybrid drive mode, and a vehicle charging mode.
8. The intelligent energy management method according to claim 1, wherein the instantaneous optimization of the engine and motor power is specifically that the energy global optimal control under the whole driving condition is converted into local optimal control in a prediction region, the running state or control parameters in the next time domain are updated through continuous rolling optimization so as to obtain an optimization result, and the high robustness of the global optimization and the real-time performance of the instantaneous optimization are combined to achieve the optimal management of the energy of the whole vehicle.
9. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to carry out the method of any one of claims 1 to 8 when executed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110001963.3A CN112668799A (en) | 2021-01-04 | 2021-01-04 | Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110001963.3A CN112668799A (en) | 2021-01-04 | 2021-01-04 | Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112668799A true CN112668799A (en) | 2021-04-16 |
Family
ID=75412636
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110001963.3A Pending CN112668799A (en) | 2021-01-04 | 2021-01-04 | Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112668799A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113247002A (en) * | 2021-06-24 | 2021-08-13 | 四川野马汽车股份有限公司 | Time-optimal-based vehicle speed control method and system |
CN113392374A (en) * | 2021-06-03 | 2021-09-14 | 联合汽车电子有限公司 | Data extraction method, vehicle service method, vehicle control system, and storage medium |
CN113486698A (en) * | 2021-04-30 | 2021-10-08 | 华中科技大学 | Identification and prediction method, storage medium and system for hydrogen fuel cell operation |
CN113479187A (en) * | 2021-07-30 | 2021-10-08 | 南京航空航天大学 | Layered different-step-length energy management method for plug-in hybrid electric vehicle |
CN113642768A (en) * | 2021-07-12 | 2021-11-12 | 南京航空航天大学 | Vehicle running energy consumption prediction method based on working condition reconstruction |
CN113788007A (en) * | 2021-11-16 | 2021-12-14 | 北京理工大学 | Layered real-time energy management method and system |
CN114137836A (en) * | 2021-11-23 | 2022-03-04 | 中国地质大学(武汉) | Vertical drilling deviation rectification control feedback correction method and device based on Gaussian mixture model |
CN114475566A (en) * | 2022-03-01 | 2022-05-13 | 重庆科技学院 | Intelligent network connection plug-in hybrid electric vehicle energy management real-time control strategy |
CN115946563A (en) * | 2023-03-13 | 2023-04-11 | 山东理工大学 | Dynamic charge stack power distribution strategy optimization method, system, terminal and medium |
CN116331184A (en) * | 2023-03-31 | 2023-06-27 | 长安大学 | Vehicle energy control method and system based on machine vision and radar information fusion |
DE102022119448A1 (en) | 2022-08-03 | 2024-02-08 | Bayerische Motoren Werke Aktiengesellschaft | Method and system for generating an energy consumption forecast for a vehicle |
WO2024073939A1 (en) * | 2022-10-08 | 2024-04-11 | 合众新能源汽车股份有限公司 | Power control method and apparatus for extended-range vehicle |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109760523A (en) * | 2019-01-21 | 2019-05-17 | 浙江工业大学 | Composite power source energy management method based on BP neural network speed prediction |
CN109895760A (en) * | 2019-04-03 | 2019-06-18 | 大连理工大学 | Plug-in hybrid-power automobile energy management method based on SOC reference locus |
CN110610260A (en) * | 2019-08-21 | 2019-12-24 | 南京航空航天大学 | Driving energy consumption prediction system, method, storage medium and equipment |
CN111080018A (en) * | 2019-12-20 | 2020-04-28 | 南京航空航天大学 | Intelligent internet automobile speed prediction method based on road traffic environment |
CN111267831A (en) * | 2020-02-28 | 2020-06-12 | 南京航空航天大学 | Hybrid vehicle intelligent time-domain-variable model prediction energy management method |
CN111815948A (en) * | 2020-06-12 | 2020-10-23 | 南京航空航天大学 | Vehicle running condition prediction method based on condition characteristics |
-
2021
- 2021-01-04 CN CN202110001963.3A patent/CN112668799A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109760523A (en) * | 2019-01-21 | 2019-05-17 | 浙江工业大学 | Composite power source energy management method based on BP neural network speed prediction |
CN109895760A (en) * | 2019-04-03 | 2019-06-18 | 大连理工大学 | Plug-in hybrid-power automobile energy management method based on SOC reference locus |
CN110610260A (en) * | 2019-08-21 | 2019-12-24 | 南京航空航天大学 | Driving energy consumption prediction system, method, storage medium and equipment |
CN111080018A (en) * | 2019-12-20 | 2020-04-28 | 南京航空航天大学 | Intelligent internet automobile speed prediction method based on road traffic environment |
CN111267831A (en) * | 2020-02-28 | 2020-06-12 | 南京航空航天大学 | Hybrid vehicle intelligent time-domain-variable model prediction energy management method |
CN111815948A (en) * | 2020-06-12 | 2020-10-23 | 南京航空航天大学 | Vehicle running condition prediction method based on condition characteristics |
Non-Patent Citations (1)
Title |
---|
丁峰等: "基于行驶工况分类的混合动力车辆速度预测方法与能量管理策略", 汽车工程, vol. 39, no. 11, 30 November 2017 (2017-11-30), pages 1223 - 1231 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113486698B (en) * | 2021-04-30 | 2023-09-26 | 华中科技大学 | Hydrogen fuel cell work identification prediction method, storage medium and system |
CN113486698A (en) * | 2021-04-30 | 2021-10-08 | 华中科技大学 | Identification and prediction method, storage medium and system for hydrogen fuel cell operation |
CN113392374A (en) * | 2021-06-03 | 2021-09-14 | 联合汽车电子有限公司 | Data extraction method, vehicle service method, vehicle control system, and storage medium |
CN113247002A (en) * | 2021-06-24 | 2021-08-13 | 四川野马汽车股份有限公司 | Time-optimal-based vehicle speed control method and system |
CN113642768A (en) * | 2021-07-12 | 2021-11-12 | 南京航空航天大学 | Vehicle running energy consumption prediction method based on working condition reconstruction |
CN113479187A (en) * | 2021-07-30 | 2021-10-08 | 南京航空航天大学 | Layered different-step-length energy management method for plug-in hybrid electric vehicle |
CN113788007A (en) * | 2021-11-16 | 2021-12-14 | 北京理工大学 | Layered real-time energy management method and system |
CN114137836A (en) * | 2021-11-23 | 2022-03-04 | 中国地质大学(武汉) | Vertical drilling deviation rectification control feedback correction method and device based on Gaussian mixture model |
CN114475566A (en) * | 2022-03-01 | 2022-05-13 | 重庆科技学院 | Intelligent network connection plug-in hybrid electric vehicle energy management real-time control strategy |
CN114475566B (en) * | 2022-03-01 | 2024-01-30 | 重庆科技学院 | Intelligent network allies oneself with inserts electric hybrid vehicle energy management real-time control strategy |
DE102022119448A1 (en) | 2022-08-03 | 2024-02-08 | Bayerische Motoren Werke Aktiengesellschaft | Method and system for generating an energy consumption forecast for a vehicle |
WO2024073939A1 (en) * | 2022-10-08 | 2024-04-11 | 合众新能源汽车股份有限公司 | Power control method and apparatus for extended-range vehicle |
CN115946563A (en) * | 2023-03-13 | 2023-04-11 | 山东理工大学 | Dynamic charge stack power distribution strategy optimization method, system, terminal and medium |
CN116331184A (en) * | 2023-03-31 | 2023-06-27 | 长安大学 | Vehicle energy control method and system based on machine vision and radar information fusion |
CN116331184B (en) * | 2023-03-31 | 2024-04-02 | 长安大学 | Vehicle energy control method and system based on machine vision and radar information fusion |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112668799A (en) | Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data | |
CN111267831B (en) | Intelligent time-domain-variable model prediction energy management method for hybrid electric vehicle | |
WO2021103625A1 (en) | Short-term vehicle speed condition real-time prediction method based on interaction between vehicle ahead and current vehicle | |
Zhou et al. | A novel energy management strategy of hybrid electric vehicle via an improved TD3 deep reinforcement learning | |
CN110775065B (en) | Hybrid electric vehicle battery life prediction method based on working condition recognition | |
Yuan et al. | Intelligent energy management strategy based on hierarchical approximate global optimization for plug-in fuel cell hybrid electric vehicles | |
CN110341690B (en) | PHEV energy management method based on deterministic strategy gradient learning | |
CN109895760B (en) | Plug-in hybrid electric vehicle energy management method based on SOC reference track | |
CN107688343B (en) | Energy control method of hybrid power vehicle | |
Khayyam et al. | Adaptive intelligent energy management system of plug-in hybrid electric vehicle | |
Singh et al. | Fuzzy logic and Elman neural network tuned energy management strategies for a power-split HEVs | |
Zhang et al. | A deep reinforcement learning-based energy management framework with Lagrangian relaxation for plug-in hybrid electric vehicle | |
CN110406526A (en) | Parallel hybrid electric energy management method based on adaptive Dynamic Programming | |
CN113479186B (en) | Energy management strategy optimization method for hybrid electric vehicle | |
CN113741199B (en) | Whole vehicle economical speed planning method based on intelligent network connection information | |
CN113135113B (en) | Global SOC (System on chip) planning method and device | |
CN115107733A (en) | Energy management method and system for hybrid electric vehicle | |
CN111923897A (en) | Intelligent energy management method for plug-in hybrid electric vehicle | |
Pozzi et al. | Ecological velocity planning through signalized intersections: A deep reinforcement learning approach | |
CN116070783B (en) | Learning type energy management method of hybrid transmission system under commute section | |
Vignesh et al. | Intelligent energy management through neuro-fuzzy based adaptive ECMS approach for an optimal battery utilization in plugin parallel hybrid electric vehicle | |
Montazeri-Gh et al. | Driving condition recognition for genetic-fuzzy HEV control | |
CN113815437A (en) | Predictive energy management method for fuel cell hybrid electric vehicle | |
Peng et al. | Ecological Driving Framework of Hybrid Electric Vehicle Based on Heterogeneous Multi-Agent Deep Reinforcement Learning | |
CN113246958A (en) | TD 3-based multi-target HEV energy management method and system |
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 |