CN102729987B - Hybrid bus energy management method - Google Patents

Abstract

The invention discloses a hybrid bus energy management method, which comprises the following steps of: acquiring a speed transition probability model according to historical data of a bus on the same line, current speed information and bus position information by a bus control unit, then estimating the speed, and optimizing energy distribution of the hybrid bus in a prediction zone according to the estimated speed to acquire current optimal power distribution ratio of a fuel oil engine to a motor; and finally determining the actual oil injection quantity and the power provided by the motor, and sending message to a motor control unit and an engine control unit through a controller area network (CAN) bus. Energy management of the hybrid bus is optimized by fully using the characteristics of the bus on the same line; and the method has the characteristics of reasonable energy distribution, high economical efficiency of fuel oil, low exhaust emission, high robustness, energy conservation and environment friendliness.

Description

A kind of hybrid-power bus energy management method

Technical field

The present invention relates to hybrid power management control technique field, be specifically related to a kind of hybrid-power bus energy management method.

Background technology

The development of auto-industry is accompanied by problem of environmental pollution and energy scarcity problem.In order to alleviate these problems, new-energy automobile is a kind of actv. approach.Because a lot of technology that pure electric automobile is relevant are also immature, there is the bottlenecks such as energy content of battery density and large electric capacity safety, current want to promote also premature.And hybrid electric vehicle (HEV) can design based on existing fuel-engined vehicle, therefore have more practical significance.This patent is exactly mainly to launch around hybrid electric vehicle.Hybrid vehicle has adopted at least two kinds of propulsions source, therefore can allow engine operation at more excellent operation interval, reaches objects such as reducing fuel oil consumption, minimizing pollutant emission, recycling braking kinetic energy.

City bus, as city main traffic instrument, has a fairly large number of feature, larger for the pollution effect of urban air, so people start to reduce the aerial contamination in city at city bus field interpolation Technology of Hybrid Electric Vehicle.At present the energy management strategy of hybrid-power bus is based on some logics judgements, however the specific aim of most of control policies a little less than, do not consider the operation characteristic of every circuit, the control effect obtaining is like this not very good.Some researchers formulate energy management strategy by means such as dynamic programmings, and this way can utilize state of cyclic operation to estimate the power demand of each time period well, reaches the object of global optimization.Yet this strategy does not have universality, while only having bus to operate according to model track definitely, just can reach control effect.But the operation that actual conditions are buses has very strong randomness, the path motion in the time of hardly can be according to custom strategies, and each consuming time be also different, if control policy probably cannot reach the control effect of expection based on fixed model.Some researchists have built probabilistic model to the operation conditions of vehicle, then adopt stochastic dynamic programming to solve, and the time complexity owing to calculating, is not adapted at optimizing in real-time vehicle operating.

Summary of the invention

The object of the invention is to overcome the deficiencies in the prior art, a kind of hybrid-power bus energy management method is provided, and the method energy of the present invention distributes rationally, fuel economy is high, robustness good, solve efficient simultaneously, be adapted at implementing in actual vehicle operational process.

The object of the invention is to be achieved through the following technical solutions: a kind of hybrid-power bus energy management method, hybrid-power bus has accelerator pedal position sensor, car speed sensor, GPS module, control unit for vehicle, control unit of engine, motor control unit and battery management unit; Accelerator pedal position sensor, car speed sensor and GPS module are all connected with control unit for vehicle, and control unit for vehicle, control unit of engine, motor control unit and battery management unit are connected by CAN bus; The method comprises the steps:

(1) car speed sensor gathers current speed information, and GPS module is obtained current vehicle location information, and the pedal position sensor of throttle obtains the location information of current throttle pedal, and battery management unit is estimated current battery charge state; The location information of current speed information, vehicle position information, Das Gaspedal and battery charge state all transfer to control unit for vehicle;

(2) after calculating, the current vehicle speed information that control unit for vehicle obtains according to speed of a motor vehicle transition probability model that early stage, historical data obtained and step 1 and vehicle position information estimate the speed of a motor vehicle;

(3) to hybrid electric vehicle, the energy distribution in prediction section is optimized the speed of a motor vehicle of estimating that control unit for vehicle obtains according to step 2, obtains the power-division ratios of fuel engines and the motor of current optimum;

(4) according to the location information of the Das Gaspedal gathering in the power-division ratios obtaining in step 3 and step 1, determine the power that actual fuel injection amount and motor should provide, then by CAN bus, send message to motor control unit and control unit of engine; The message that control unit of engine sends over according to reception control unit for vehicle regulates the horsepower output of driving engine; The message that motor control unit sends over according to reception control unit for vehicle is controlled the horsepower output of electrical motor.

Further, described step 2 can specifically be divided into several sub-steps below:

(2.1) the historical speed of a motor vehicle in the operation of same section according to bus, calculate bus at the speed of a motor vehicle transition probability matrix of each displacement point, and build speed of a motor vehicle transition probability model: first by the spatial discretization of speed of a motor vehicle value, 0 to 60km/h speed interval is divided into and take the 25 lattice speed of a motor vehicle values that 2.5km/h is interval; Then according to historical speed of a motor vehicle path, each position on vehicle operating path trains respectively speed of a motor vehicle transition probability matrix, and these speed of a motor vehicle transition probability matrixs are stored in control unit for vehicle; Relevant to the value quantity of state variable after discretization in the dimension of the state transition probability matrix P at S place, position (S), if having 25 possible values after speed scattering, the matrix that P (S) is 25 * 25; In P (S) matrix, the probability of b speed scattering value is transferred in the element representative of the capable b row of a from a speed scattering value, and a and b are the natural number of 1-25; In P (S) matrix, the method for calculating of the element P (S, a, b) of the capable b row of a is as follows:

$P(S,a,b)=\frac{N_b(S,a)}{N_a\left(S\right)};$

Wherein, N _a(S) be the number that in historical path, the speed of a motor vehicle at S place, position is a discrete value; N _b(S, a) for the speed of a motor vehicle at S place, position in historical path is the sum that the speed of a motor vehicle of a discrete value at S+1 place, position changes b discrete value into.From all state transition probability matrixs of the origin-to-destination of Vehicle Driving Cycle, formed speed of a motor vehicle transition probability model;

(2.2) after calculating, the current vehicle speed information that control unit for vehicle obtains according to above-mentioned speed of a motor vehicle transition probability model and step 1 and vehicle position information estimate the speed of a motor vehicle: first current vehicle speed v (S) is converted into 1 * 25 vectorial form, the 1st probability that element is 0 corresponding to the speed of a motor vehicle in this vector, n element is the probability of (n-1) * 2.5km/h corresponding to the speed of a motor vehicle, the natural number that n is 1-25; V (S) multiplies each other to obtain with speed of a motor vehicle transition probability model and at current location S, from current vehicle speed v (S), changes all probable values of next speed of a motor vehicle v (S+1) into:

v(S+1)＝v(S)×P(S)；

Above formula acquired results v (S+1) is also the vector of 1 * 25, then just obtained position S+1 likely the set of the speed of a motor vehicle for { v _j(S+1) | j=1,2 ..., N _s+1, wherein, N _s+1be illustrated in the sum of the non-vanishing possible speed of a motor vehicle of S+1 location probability, the speed of a motor vehicle v that each is possible _j(S+1) corresponding probability is P _j(S), P _j(S) value is exactly j non-vanishing value in vector v (S+1); By each possible speed of a motor vehicle, continued to obtain the follow-up speed of a motor vehicle and probability according to speed of a motor vehicle transition probability model recursion afterwards, in order to simplify follow-up computation complexity, in S+2 step and estimation range afterwards, directly use the speed of a motor vehicle path trajectory (j) of maximum probability as v _j(S+1) follow-up speed of a motor vehicle prediction path, next step branch value of several speed of a motor vehicle and follow-up maximum possible path are combined into estimating of the speed of a motor vehicle.

Further, the objective function in described step 3, the energy distribution of hybrid electric vehicle being optimized is:

Wherein, u is controlling quantity, u (i) represents that i section moves the ratio of middle engine power and total power demand, F (i, u (i)) represent i section move in fuel oil consumption under the effect of controlling quantity u (i), Δ SOC (i, u (i)) represent i section move in battery charge state reduction under the effect of controlling quantity u (i), P is the length (being the total prediction step number of estimating the speed of a motor vehicle obtaining in step 2) of forecast interval, G is weights, and E is the symbol of asking for expectation value; Target is to ask for an optimal control sequence (being the power-division ratios of each position in estimation range) in prediction time domain; The circular of above-mentioned F (i, u (i)) is:

F(i,u(i))＝map_f(Treq(i)×u(i),ω _ICE(i))Δt(i)；

Wherein, map_f is the bivariate table of a fuel consumption rate, and the concrete data in this bivariate table are relevant to driving engine, T _reqfor the torque of the required output of torsion coupler, ω _iCEfor the rotating speed of driving engine, T _reqwith ω _iCEcan obtain according to automobile longitudinal kinetics equation, Δ t (i) represents that vehicle moves the time of consumption at i section.

The circular of Δ SOC (i, u (i)) is:

$\mathrm{\ΔSOC}(i,u\left(i\right))=[(U-\sqrt{U^2-{4P}_{\mathrm{req}}(1-u\left(i\right))R})/2R]\mathrm{\Δt}\left(i\right)/Q;$

Wherein, P _reqfor total demand power, the equivalent open-loop voltage that U is battery, the equivalent internal resistance that R is battery, total electric weight that Q is battery.

Initial speed of a motor vehicle during optimization and initial state-of-charge are all taken from the value that step 1 collects.Concrete optimization method is divided into two stages: in first stage, only consider the independent drive pattern of driving engine, the independent drive pattern of electrical motor, these three kinds of mode of operations of combination drive pattern; Power-division ratios under these three kinds of mode of operations (ratio of engine power and total power demand) is less than or equal to 1; Then by power-division ratios discretization in current feasible zone; In forecast interval, the power-division ratios of all positions is first initially 1, then tastes the power-division ratios of each point is lowered to next possible values, then for each point calculates, adjusts the impact bringing, and impact specifically represents with following decision variable:

k＝∑P _j[(Δf _j-GΔSOC _j)/ΔSOC _j]＝∑P _j[Δf _j/ΔSOC _j-G]；

Wherein, P _jbe illustrated in the probability of transferring to j bar prediction curve in speed of a motor vehicle prediction model, Δ f _jbe illustrated under the operating mode of j bar prediction curve the oil consumption reduction that Modulating Power distribution ratio causes, Δ SOC _jrepresent the state-of-charge reducing; Whether decision variable k can be used for judging at current point is worth cost electric energy to reduce oil consumption, k is illustrated in more greatly this point, and to consume the benefit that identical electric energy brings larger, so calculate after the decision variable of each point in forecast interval, select that point of the value maximum of k and lower power-division ratios, above-mentioned is the searching process of an iteration, each optimizing is all to select the maximum position of k value to lower power-division ratios, and the end condition of iteration is that k=0 or SOC reach lower limit; Just enter afterwards the optimizing process in next stage; The mode of operation that the optimization of second stage is considered is power generation mode or maintains the last pattern of first stage.After first stage optimization, still those points in the independent drive pattern of driving engine just likely change power generation mode in subordinate phase; The position candidate of generating using those as subordinate phase.Obtain, after the position candidate of generating, will determining optimum electric energy generated.The optimizing process of subordinate phase is specially: first taste the power-division ratios of each point is risen to next possible values, power-division ratios (ratio of engine power and total power demand) is greater than at 1 o'clock and then for each point calculates, adjusts the impact bringing, and affects and specifically with following decision variable, represents:

k'＝∑P _j[(GΔSOC' _j-Δf _j')/Δf _j']＝∑P _j[GΔSOC _j'/Δf _j'-1]；

Wherein, P _jbe illustrated in the probability of transferring to j bar prediction curve in speed of a motor vehicle prediction model, Δ f _j' be illustrated under the operating mode of j bar prediction curve for the extra fuel oil consumption increasing that generates electricity, Δ SOC ' _jrepresent corresponding SOC increment; Whether decision variable k ' can be used for judging at current point is worth the extra fuel oil of cost to produce electric energy; From candidate point, select the some regulating power distribution ratio of the value maximum of k '; Then enter next iteration, again calculate the value of k ', and in the maximum position adjustments power-division ratios of k '; The position of generating is all the position that cost performance is the highest like this; Through the optimization in two stages above, in forecast interval, the power-division ratios of each position has been set to optimum position.

The present invention has following technique effect: the present invention makes full use of bus in the feature of same section operation, and the energy management of hybrid-power bus is optimized.The historical speed of a motor vehicle-the displacement relation that the present invention is based on statistics obtains the speed of a motor vehicle of estimating after current displacement point, and carrying out optimizing power distribution ratio according to the speed of a motor vehicle of estimating, this invention has advantages of that the energy distributes rationally, fuel economy is high, exhaust emissions is few, robustness is good, energy-conserving and environment-protective.

Accompanying drawing explanation

Fig. 1 is the implementing procedure schematic diagram of the embodiment of the present invention;

Fig. 2 is the framed structure schematic diagram of the embodiment of the present invention;

Fig. 3 is that the speed of a motor vehicle of the present invention is estimated schematic diagram;

Fig. 4 is the diagram of circuit of the first stage of the present invention while optimizing;

Fig. 5 is the diagram of circuit of the subordinate phase of the present invention while optimizing.

The specific embodiment

In hybrid-power bus, there is accelerator pedal position sensor, car speed sensor, GPS module, control unit for vehicle, control unit of engine, motor control unit and battery management unit, accelerator pedal position sensor, car speed sensor and GPS module are all connected with control unit for vehicle, and control unit for vehicle, control unit of engine, motor control unit and battery management unit are connected by CAN bus.Wherein, control unit for vehicle receives speed information by accelerator pedal position sensor and car speed sensor and accelerator pedal position information draws optimum power-division ratios, then passes to control unit of engine, motor control unit and battery management unit; The variablees such as throttle opening are controlled in the instruction that control unit of engine sends over according to the control unit for vehicle receiving, thereby regulate engine output; The horsepower output of electrical motor is controlled in the instruction that motor control unit sends over according to the control unit for vehicle receiving; The state-of-charge that battery management unit estimating battery is current is also passed to automobile control unit.

As shown in Figure 1, a kind of hybrid-power bus energy management method of the embodiment of the present invention, implementation step is as follows:

1, car speed sensor gathers current speed information, and GPS module is obtained current vehicle location information, and the pedal position sensor of throttle obtains the location information of current throttle pedal, and battery management unit is estimated current battery charge state; Location information and the battery charge state of current speed information, vehicle position information, Das Gaspedal transfer to control unit for vehicle.

2, after calculating, the current vehicle speed information that control unit for vehicle obtains according to speed of a motor vehicle transition probability model that early stage, historical data obtained and step 1 and vehicle position information estimate the speed of a motor vehicle.

Step 2 can specifically be divided into several sub-steps below:

2.1, the historical speed of a motor vehicle in the operation of same section according to bus, calculates bus at the speed of a motor vehicle transition probability matrix of each displacement point, and builds speed of a motor vehicle transition probability model.Off-line modeling is first by the spatial discretization of speed of a motor vehicle value, and 0 to 60km/h speed interval is divided into and take the 25 lattice speed of a motor vehicle values that 2.5km/h is interval.Then according to historical speed of a motor vehicle path, each position on vehicle operating path trains respectively speed of a motor vehicle transition probability matrix, and these speed of a motor vehicle transition probability matrixs are stored in control unit for vehicle.Relevant to the value quantity of state variable after discretization in the dimension of the state transition probability matrix P at S place, position (S), if having 25 possible values after speed scattering, the matrix that P (S) is 25 * 25.In P (S) matrix, the probability of b speed scattering value is transferred in the element representative of the capable b row of a from a speed scattering value, and concrete probable value is by various may the acquisition of the speed of a motor vehicle variation at same position place in statistical history speed of a motor vehicle path.The state transition probability matrix of each position should be different.For example the starting stage, there is a strong possibility for the next speed of a motor vehicle higher than current vehicle speed; The probability that the speed of a motor vehicle declines when soon running into crossing is larger.From all state transition probability matrixs of the origin-to-destination of Vehicle Driving Cycle, formed speed of a motor vehicle transition probability model, a and b are the natural number of 1-25.In P (S) matrix, the method for calculating of the element P (S, a, b) of the capable b row of a is as follows:

$P(S,a,b)=\frac{N_b(S,a)}{N_a\left(S\right)};$

Wherein, N _a(S) be the number that in historical path, the speed of a motor vehicle at S place, position is a discrete value; N _b(S, a) for the speed of a motor vehicle at S place, position in historical path is the sum that the speed of a motor vehicle of a discrete value at S+1 place, position changes b discrete value into.From all state transition probability matrixs of the origin-to-destination of Vehicle Driving Cycle, formed speed of a motor vehicle transition probability model;

2.2, after calculating, the current vehicle speed information that control unit for vehicle obtains according to above-mentioned speed of a motor vehicle transition probability model and step 1 and vehicle position information estimate the speed of a motor vehicle.

It is bus current vehicle speed and current location that the data that the speed of a motor vehicle need to input are estimated in described calculating, then according to step 2) in speed of a motor vehicle transition probability model, output estimating the following speed of a motor vehicle.Concrete method of calculating is: first current vehicle speed v (S) is converted into 1 * 25 vectorial form, the 1st probability that element is 0 corresponding to the speed of a motor vehicle in this vector, n element is the probability of (n-1) * 2.5km/h corresponding to the speed of a motor vehicle, the natural number that n is 1-25.V (S) multiplies each other to obtain with speed of a motor vehicle transition probability model and at current location S, from current vehicle speed v (S), changes all probable values of next speed of a motor vehicle v (S+1) into:

v(S+1)＝v(S)×P(S)；

Above formula acquired results v (S+1) is also the vector of a 1 * n, then just obtained position S+1 likely the set of the speed of a motor vehicle for { v _j(S+1) | j=1,2 ..., N _s+1, wherein, N _s+1be illustrated in the sum of the non-vanishing possible speed of a motor vehicle of S+1 location probability, the speed of a motor vehicle v that each is possible _j(S+1) corresponding probability is P _j(S), P _j(S) value is exactly j non-vanishing value in vector v (S+1).

By each possible speed of a motor vehicle, continued to obtain the follow-up speed of a motor vehicle and probability according to speed of a motor vehicle transition probability model recursion afterwards, in order to simplify follow-up computation complexity, in S+2 step and estimation range afterwards, directly use the speed of a motor vehicle path trajectory (j) of maximum probability as v _j(S+1) follow-up speed of a motor vehicle prediction path, next step branch value of several speed of a motor vehicle and follow-up maximum possible path are combined into estimating of the speed of a motor vehicle, roughly as shown in Figure 3.

3, to hybrid electric vehicle, the energy distribution in prediction section is optimized the speed of a motor vehicle of estimating that control unit for vehicle obtains according to step 2, obtains the power-division ratios of fuel engines and the motor of current optimum.

The objective function in described step 3, the energy distribution of hybrid electric vehicle being optimized is:

Wherein, u is controlling quantity, u (i) represents that i section moves the ratio of middle engine power and total power demand, F (i, u (i)) represent i section move in fuel oil consumption under the effect of controlling quantity u (i), Δ SOC (i, u (i)) represent i section move in battery charge state reduction under the effect of controlling quantity u (i), P is the length (being the total prediction step number of estimating the speed of a motor vehicle obtaining in step 2) of forecast interval, G is weights, and E is the symbol of asking for expectation value.Target is to ask for an optimal control sequence (being the power-division ratios of each position in estimation range) in prediction time domain.The circular of above-mentioned F (i, u (i)) is:

F(i,u(i))＝map_f(Treq(i)×u(i),ω _ICE(i))Δt(i)；

Wherein, map_f is the bivariate table of a fuel consumption rate, and the concrete data in this bivariate table are relevant to driving engine, T _reqfor the torque of the required output of torsion coupler, ω _iCEfor the rotating speed of driving engine, T _reqwith ω _iCEcan obtain according to automobile longitudinal kinetics equation, Δ t (i) represents that vehicle moves the time of consumption at i section.

The circular of Δ SOC (i, u (i)) is:

$\mathrm{\ΔSOC}(i,u\left(i\right))=[(U-\sqrt{U^2-{4P}_{\mathrm{req}}(1-u\left(i\right))R})/2R]\mathrm{\Δt}\left(i\right)/Q;$

Wherein, P _reqfor total demand power, the equivalent open-loop voltage that U is battery, the equivalent internal resistance that R is battery, total electric weight that Q is battery.

Initial speed of a motor vehicle during optimization and initial state-of-charge are all taken from the value that step 1 collects.Concrete optimization method is divided into two stages: the Optimizing Flow of first stage as shown in Figure 4, is only considered the independent drive pattern of driving engine, the independent drive pattern of electrical motor, these three kinds of mode of operations of combination drive pattern in first stage.Power-division ratios under these three kinds of mode of operations (ratio of engine power and total power demand) is less than or equal to 1.Then by power-division ratios discretization in current feasible zone.In forecast interval, the power-division ratios of all positions is first initially 1, then tastes the power-division ratios of each point is lowered to next possible values, then for each point calculates, adjusts the impact bringing, and impact specifically represents with following decision variable:

k＝∑P _j[(Δf _j-GΔSOC _j)/ΔSOC _j]＝∑P _j[Δf _j/ΔSOC _j-G]；

Wherein, P _jbe illustrated in the probability of transferring to j bar prediction curve in speed of a motor vehicle prediction model, Δ f _jbe illustrated under the operating mode of j bar prediction curve the oil consumption reduction that Modulating Power distribution ratio causes, Δ SOC _jrepresent the state-of-charge reducing.Whether decision variable k can be used for judging at current point is worth cost electric energy to reduce oil consumption.K is illustrated in more greatly this point, and to consume the benefit that identical electric energy brings larger.So calculate after the decision variable of each point in forecast interval, select that point of the value maximum of k and lower power-division ratios.Above-mentioned is the searching process of an iteration, and each optimizing is all to select the maximum position of k value to lower power-division ratios.The end condition of iteration is that k=0 or SOC reach lower limit.Just enter afterwards the optimizing process in next stage.The mode of operation that the optimization of second stage is considered is power generation mode or maintains the last pattern of first stage.After first stage optimization, still those points in the independent drive pattern of driving engine just likely change power generation mode in subordinate phase; The position candidate of generating using those as subordinate phase.Obtain, after the position candidate of generating, will determining optimum electric energy generated.The optimizing process of subordinate phase as shown in Figure 5, process and the first stage of iteration optimizing are somewhat similar, first taste the power-division ratios of each point is risen to next possible values, power-division ratios (ratio of engine power and total power demand) is greater than at 1 o'clock and then for each point calculates, adjusts the impact bringing, and affects and specifically with following decision variable, represents:

k'＝∑P _j[(GΔSOC' _j-Δf _j')/Δf _j']＝∑P _j[GΔSOC _j'/Δf _j'-1]；

Wherein, P _jbe illustrated in the probability of transferring to j bar prediction curve in speed of a motor vehicle prediction model, Δ f _j' be illustrated under the operating mode of j bar prediction curve for the extra fuel oil consumption increasing that generates electricity, Δ SOC ' _jrepresent corresponding SOC increment.Whether decision variable k ' can be used for judging at current point is worth the extra fuel oil of cost to produce electric energy.From candidate point, select the some regulating power distribution ratio of the value maximum of k '.Then enter next iteration, again calculate the value of k ', and in the maximum position adjustments power-division ratios of k '.The position of generating is all the position that cost performance is the highest like this.How to use the electric energy of new generation to be based on the optimum results of first stage, if the optimization of first stage is due to k<0 while stopping, the so new electric energy producing can not reallocated, if first stage optimization is because SOC is lower than lower limit while stopping, the so new electric energy producing can be used in the position of the value maximum of k, then the value of upgrading k enters next iteration, and the end condition of subordinate phase Optimized Iterative is k ' <0.Through the optimization in two stages above, in forecast interval, the power-division ratios of each position has been set to optimum position.

4, according to the location information of the Das Gaspedal gathering in the power-division ratios obtaining in step 3 and step 1, determine the power that actual fuel injection amount and motor should provide, then by CAN bus, send corresponding message to motor control unit and control unit of engine, concrete message form is relevant to the a2l file of each automobile vendor.The peripheral relevant framed structure schematic diagram of whole automobile control unit as shown in Figure 2.Control unit of engine is responsible for receiving the instruction that control unit for vehicle sends over and is controlled the effect that the variablees such as throttle opening reach adjusting power of engine output; Motor control unit is used for controlling the power stage of electrical motor.

When vehicle operating can be upgraded the present speed of a motor vehicle, position and battery charge state SOC during to next displaced segments, follow estimating according to the new state information updating speed of a motor vehicle, and the power-division ratios in ensuing forecast interval is optimized, and the horsepower output of control engine and motor correspondingly.

The present invention has considered statistical law and the randomness that bus moves on same circuit, the control algorithm adopting changes traditional multi-modal problem into two class subproblems, the thought of dividing and ruling embodying, and what utilize is the thought of greedy algorithm, greatly accelerated the speed that solves of optimization problem when determining electric energy generated and power consumption.Be applicable to vehicle is optimized in real time execution.

Claims (3)

1. a hybrid-power bus energy management method, hybrid-power bus has accelerator pedal position sensor, car speed sensor, GPS module, control unit for vehicle, control unit of engine, motor control unit and battery management unit; Accelerator pedal position sensor, car speed sensor and GPS module are all connected with control unit for vehicle, and control unit for vehicle, control unit of engine, motor control unit and battery management unit are connected by CAN bus; It is characterized in that, the method comprises the steps:

(1) car speed sensor gathers current speed information, and GPS module is obtained current vehicle location information, and the pedal position sensor of throttle obtains the location information of current throttle pedal, and battery management unit is estimated current battery charge state; The location information of current speed information, vehicle position information, Das Gaspedal and battery charge state all transfer to control unit for vehicle;

(2) after calculating, the current vehicle speed information that control unit for vehicle obtains according to speed of a motor vehicle transition probability model that early stage, historical data obtained and step 1 and vehicle position information estimate the speed of a motor vehicle;

(3) to hybrid electric vehicle, the energy distribution in prediction section is optimized the speed of a motor vehicle of estimating that control unit for vehicle obtains according to step 2, obtains the power-division ratios of fuel engines and the motor of current optimum;

(4) according to the location information of the Das Gaspedal gathering in the power-division ratios obtaining in step 3 and step 1, determine the power that actual fuel injection amount and motor should provide, then by CAN bus, send message to motor control unit and control unit of engine; The message that control unit of engine sends over according to reception control unit for vehicle regulates the horsepower output of driving engine; The message that motor control unit sends over according to reception control unit for vehicle is controlled the horsepower output of electrical motor.

2. hybrid-power bus energy management method according to claim 1, is characterized in that, described step (2) can specifically be divided into several sub-steps below:

(2.1) the historical speed of a motor vehicle in the operation of same section according to bus, calculate bus at the speed of a motor vehicle transition probability matrix of each displacement point, and build speed of a motor vehicle transition probability model: first by the spatial discretization of speed of a motor vehicle value, 0 to 60km/h speed interval is divided into and take the 25 lattice speed of a motor vehicle values that 2.5km/h is interval; Then according to historical speed of a motor vehicle path, each position on vehicle operating path trains respectively speed of a motor vehicle transition probability matrix, and these speed of a motor vehicle transition probability matrixs are stored in control unit for vehicle; Relevant to the value quantity of state variable after discretization in the dimension of the state transition probability matrix P at S place, position (S), if having 25 possible values after speed scattering, the matrix that P (S) is 25 * 25; In P (S) matrix, the probability of b speed scattering value is transferred in the element representative of the capable b row of a from a speed scattering value, and a and b are the natural number of 1-25; In P (S) matrix, the method for calculating of the element P (S, a, b) of the capable b row of a is as follows:

$P(S,a,b)=\frac{N_b(S,a)}{N_a\left(S\right)};$

Wherein, N _a(S) be the number that in historical path, the speed of a motor vehicle at S place, position is a discrete value; N _b(S, a) for the speed of a motor vehicle at S place, position in historical path is the sum that the speed of a motor vehicle of a discrete value at S+1 place, position changes b discrete value into; From all state transition probability matrixs of the origin-to-destination of Vehicle Driving Cycle, formed speed of a motor vehicle transition probability model;

(2.2) after calculating, the current vehicle speed information that control unit for vehicle obtains according to above-mentioned speed of a motor vehicle transition probability model and step (1) and vehicle position information estimate the speed of a motor vehicle: first current vehicle speed v (S) is converted into 1 * 25 vectorial form, the 1st probability that element is 0 corresponding to the speed of a motor vehicle in this vector, n element is the probability of (n-1) * 2.5km/h corresponding to the speed of a motor vehicle, the natural number that n is 1-25; V (S) multiplies each other to obtain with speed of a motor vehicle transition probability model and at current location S, from current vehicle speed v (S), changes all probable values of next speed of a motor vehicle v (S+1) into:

v(S+1)＝v(S)×P(S)；

Above formula acquired results v (S+1) is also the vector of 1 * 25, then just obtained position S+1 likely the set of the speed of a motor vehicle for { v _j(S+1) | j=1,2 ..., N _s+1, wherein, N _s+1be illustrated in the sum of the non-vanishing possible speed of a motor vehicle of S+1 location probability, the speed of a motor vehicle v that each is possible _j(S+1) corresponding probability is P _j(S), P _j(S) value is exactly j non-vanishing value in vector v (S+1); By each possible speed of a motor vehicle, continued to obtain the follow-up speed of a motor vehicle and probability according to speed of a motor vehicle transition probability model recursion afterwards, in order to simplify follow-up computation complexity, in S+2 step and estimation range afterwards, directly use the speed of a motor vehicle path trajectory (j) of maximum probability as v _j(S+1) follow-up speed of a motor vehicle prediction path, next step branch value of several speed of a motor vehicle and follow-up maximum possible path are combined into estimating of the speed of a motor vehicle.

3. hybrid-power bus energy management method according to claim 1, is characterized in that, the objective function in described step (3), the energy distribution of hybrid electric vehicle being optimized is:

Wherein, u is controlling quantity, u (i) represents that i section moves the ratio of middle engine power and total power demand, F (i, u (i)) represent i section move in fuel oil consumption under the effect of controlling quantity u (i), Δ SOC (i, u (i)) represent i section move in battery charge state reduction under the effect of controlling quantity u (i), P is the length (being the total prediction step number of estimating the speed of a motor vehicle obtaining in step (2)) of forecast interval, G is weights, and E is the symbol of asking for expectation value; Target is to ask for an optimal control sequence (being the power-division ratios of each position in estimation range) in prediction time domain; The circular of above-mentioned F (i, u (i)) is:

F(i,u(i))＝map_f(Treq(i)×u(i),ω _ICE(i))Δt(i)；

Wherein, map_f is the bivariate table of a fuel consumption rate, and the concrete data in this bivariate table are relevant to driving engine, T _reqfor the torque of the required output of torsion coupler, ω _iCEfor the rotating speed of driving engine, T _reqwith ω _iCEcan obtain according to automobile longitudinal kinetics equation, Δ t (i) represents that vehicle moves the time of consumption at i section;

The circular of Δ SOC (i, u (i)) is:

$\mathrm{\&Delta;SOC}(i,u\left(i\right))=[(U-\sqrt{U^2-{4P}_{\mathrm{req}}(1-u\left(i\right))R})/2R]\mathrm{\&Delta;t}\left(i\right)/Q;$

Wherein, P _reqfor total demand power, the equivalent open-loop voltage that U is battery, the equivalent internal resistance that R is battery, total electric weight that Q is battery;

Initial speed of a motor vehicle during optimization and initial state-of-charge are all taken from the value that step (1) collects; Concrete optimization method is divided into two stages: in first stage, only consider the independent drive pattern of driving engine, the independent drive pattern of electrical motor, these three kinds of mode of operations of combination drive pattern; Power-division ratios under these three kinds of mode of operations (ratio of engine power and total power demand) is less than or equal to 1; Then by power-division ratios discretization in current feasible zone; In forecast interval, the power-division ratios of all positions is first initially 1, then tastes the power-division ratios of each point is lowered to next possible values, then for each point calculates, adjusts the impact bringing, and impact specifically represents with following decision variable:

k＝∑P _j[(Δf _j-GΔSOC _j)/ΔSOC _j]＝∑P _j[Δf _j/ΔSOC _j-G]；

Wherein, P _jbe illustrated in the probability of transferring to j bar prediction curve in speed of a motor vehicle prediction model, Δ f _jbe illustrated under the operating mode of j bar prediction curve the oil consumption reduction that Modulating Power distribution ratio causes, Δ SOC _jrepresent the state-of-charge reducing; Whether decision variable k can be used for judging at current point is worth cost electric energy to reduce oil consumption, k is illustrated in more greatly this point, and to consume the benefit that identical electric energy brings larger, so calculate after the decision variable of each point in forecast interval, select that point of the value maximum of k and lower power-division ratios, above-mentioned is the searching process of an iteration, each optimizing is all to select the maximum position of k value to lower power-division ratios, and the end condition of iteration is that k=0 or SOC reach lower limit; Just enter afterwards the optimizing process in next stage; The mode of operation that the optimization of second stage is considered is power generation mode or maintains the last pattern of first stage; After first stage optimization, still those points in the independent drive pattern of driving engine just likely change power generation mode in subordinate phase; The position candidate of generating using those as subordinate phase; Obtain, after the position candidate of generating, will determining optimum electric energy generated; The optimizing process of subordinate phase is specially: first taste the power-division ratios of each point is risen to next possible values, power-division ratios (ratio of engine power and total power demand) is greater than at 1 o'clock and then for each point calculates, adjusts the impact bringing, and affects and specifically with following decision variable, represents:

k'＝∑P _j[(GΔSOC' _j-Δf _j')/Δf _j']＝∑P _j[GΔSOC _j'/Δf _j'-1]；

Wherein, P _jbe illustrated in the probability of transferring to j bar prediction curve in speed of a motor vehicle prediction model, Δ f _j' be illustrated under the operating mode of j bar prediction curve for the extra fuel oil consumption increasing that generates electricity, Δ SOC ' _jrepresent corresponding SOC increment; Whether decision variable k ' can be used for judging at current point is worth the extra fuel oil of cost to produce electric energy; From candidate point, select the some regulating power distribution ratio of the value maximum of k '; Then enter next iteration, again calculate the value of k ', and in the maximum position adjustments power-division ratios of k '; The position of generating is all the position that cost performance is the highest like this; Through the optimization in two stages above, in forecast interval, the power-division ratios of each position has been set to optimum position.