CN111695202A - Fuel cell vehicle fuzzy control strategy optimization method based on approximate model - Google Patents

Fuel cell vehicle fuzzy control strategy optimization method based on approximate model Download PDF

Info

Publication number
CN111695202A
CN111695202A CN202010568570.6A CN202010568570A CN111695202A CN 111695202 A CN111695202 A CN 111695202A CN 202010568570 A CN202010568570 A CN 202010568570A CN 111695202 A CN111695202 A CN 111695202A
Authority
CN
China
Prior art keywords
fuel cell
control strategy
design
fuzzy control
approximate model
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.)
Granted
Application number
CN202010568570.6A
Other languages
Chinese (zh)
Other versions
CN111695202B (en
Inventor
李蒙
赵震
王铁
王戎
冯凯
蔡龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Taiyuan University of Technology
Original Assignee
Taiyuan University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Taiyuan University of Technology filed Critical Taiyuan University of Technology
Priority to CN202010568570.6A priority Critical patent/CN111695202B/en
Publication of CN111695202A publication Critical patent/CN111695202A/en
Application granted granted Critical
Publication of CN111695202B publication Critical patent/CN111695202B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/15—Vehicle, aircraft or watercraft design
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
    • B60L15/00—Methods, circuits, or devices for controlling the traction-motor speed of electrically-propelled vehicles
    • B60L15/20—Methods, circuits, or devices for controlling the traction-motor speed of electrically-propelled vehicles for control of the vehicle or its driving motor to achieve a desired performance, e.g. speed, torque, programmed variation of speed
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
    • B60L58/00—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles
    • B60L58/30—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles for monitoring or controlling fuel cells
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/60—Other road transportation technologies with climate change mitigation effect
    • Y02T10/72—Electric energy management in electromobility
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02T90/40—Application of hydrogen technology to transportation, e.g. using fuel cells

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Artificial Intelligence (AREA)
  • Power Engineering (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Physiology (AREA)
  • Mathematical Physics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Sustainable Development (AREA)
  • Sustainable Energy (AREA)
  • Fuel Cell (AREA)
  • Electric Propulsion And Braking For Vehicles (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a fuel cell automobile fuzzy control strategy optimization method based on an approximate model, which introduces an ellipsoid-based neural network approximate model method into the field of fuel cell automobiles to optimize a control strategy. And establishing a complete vehicle simulation approximate model of the ellipsoid-based neural network related to the optimization variables, and optimizing related parameters of a membership function in a fuzzy control strategy by using a multi-island genetic algorithm on the basis of the approximate model and taking the minimum equivalent hydrogen consumption as a target function. The optimized fuel cell system has more stable output power, the probability of being in the optimal efficiency area of the fuel cell is increased, the economic performance of the whole vehicle is improved, and the fuel cell system has important significance for the application development of fuel cell vehicles.

Description

Fuel cell vehicle fuzzy control strategy optimization method based on approximate model
Technical Field
The invention relates to the technical field of fuel cell automobiles, in particular to a fuzzy control strategy optimization method for a whole fuel cell automobile based on an ellipsoid-based neural network approximation model.
Background
A fuel cell is a power generation device that converts chemical energy into electric energy, and can continuously and stably output electric energy to generate water and heat by using hydrogen as a fuel and air/oxygen as an oxidant without using fossil fuel. Along with the development of fuel cell technology, fuel cell vehicles are also rapidly developed, the fuel cell vehicles can realize zero pollution and zero emission, and the problems of short endurance mileage and the like of pure electric vehicles are solved.
Because the output characteristic of the fuel cell is soft and the dynamic response characteristic is poor, auxiliary energy needs to be added to carry out peak clipping and valley filling in the driving process of the automobile, so that the output power of the fuel cell is maintained in a relatively stable state, and the service life of the fuel cell is prolonged. Due to the existence of a plurality of energy sources, the management problem of the energy of the whole vehicle is very important. The fuzzy control strategy is one of the most common and practical whole vehicle control modes of the fuel cell vehicle as a regular whole vehicle energy control strategy. In the design process of the fuzzy controller, the formulation of the membership function and the fuzzy rule mainly depends on engineering and practical experience, and the membership function and the fuzzy rule have larger contingency and uncertainty, so that the performance of a fuel cell vehicle is difficult to reach the best, the economy of the whole vehicle is influenced, and a larger optimization and improvement space is provided.
Disclosure of Invention
In order to solve the problems, the invention provides a fuel cell vehicle energy fuzzy control strategy optimization method based on an approximate model. An ellipsoid-based neural network approximation model method is introduced into the field of fuel cell vehicle control strategy optimization, and sensitivity analysis of equivalent hydrogen consumption is performed on related design parameters of membership functions in a fuzzy control strategy, so that optimized design variables are screened out. And sampling the screened design variables through the optimized Latin hypercube design, and establishing an ellipsoid-based neural network whole vehicle simulation approximate model. On the basis of the model, the fuzzy control strategy is optimized by using a multi-island genetic algorithm with the minimum equivalent hydrogen consumption as an objective function.
In order to achieve the aim, the fuel cell vehicle energy fuzzy control strategy optimization method based on the approximate model comprises the following steps:
designing a fuel cell vehicle energy fuzzy control management strategy: the fuel cell vehicle Mamdani type fuzzy controller adopts a double-input single-output structural form, and two input variables of the fuzzy controller are selected as the required power P of the whole vehicleloadAnd the SOC of the power battery is relative to the set value SOC*△ SOC, the fuzzy controller output variable is the fuel cell system reference output power Pf(ii) a Defining fuzzy distribution of input and output quantities and formulating a fuzzy control rule;
establishing an ellipsoid-based neural network whole vehicle simulation approximate model related to an optimized variable:
(1) determining an optimal design variable: screening the optimized design variables of the input and output quantity membership function by calculating the parameter contribution rate, and reducing the number of the design variables; firstly, sampling design variables by adopting an optimized Latin hypercube method to complete test design; secondly, sensitivity analysis is carried out, a multiple quadratic regression model is established according to the sample points and the corresponding simulation output results y,
Figure BDA0002548458240000021
wherein, β0、βi、βijCoefficients of the regression model; normalizing input design variables to [ -1, 1]Fitting the multiple quadratic regression model by using a least square method to obtain a model coefficient SiWill SiConverting into percentage of contribution rate to the equivalent hydrogen consumption, and reflecting the contribution of each design variable to the response-the equivalent hydrogen consumption; determining a fuzzy control strategy design variable which needs to be optimized finally according to the influence result of the design variable on the equivalent hydrogen consumption;
(2) establishing an overall vehicle simulation approximate model based on an ellipsoid-based neural network:
sampling the screened design variables to be optimized finally by adopting an optimized Latin hypercube design method, and training an ellipsoid-based neural network by utilizing the obtained sample data, thereby establishing a complete vehicle simulation approximate model about the optimized design variables;
based on a complete vehicle simulation approximate model, a multi-island genetic algorithm is utilized to optimize a fuzzy control strategy: optimizing a fuzzy control strategy by adopting a multi-island genetic algorithm to ensure that the fuel cell automobile realizes the optimal economy of the whole automobile; the optimization problem is described by the following formula:
Figure BDA0002548458240000022
wherein the content of the first and second substances,
Figure BDA0002548458240000023
for the equivalent hydrogen consumption of the vehicle, i.e. the objective function in the optimization process, to achieve optimum economy of the entire vehicle travel, the target value needs to be minimized, gi(X) is a constraint (design space constraint) in the optimization process. And (3) redesigning a fuzzy control strategy according to the optimized variable value, and completing the optimization of the fuzzy control strategy of the fuel cell vehicle based on the approximate model, wherein the output power of the fuel cell system after the optimization is more stable than that before the optimization, the equivalent hydrogen consumption is reduced compared with that before the optimization, the equivalent hundred kilometer hydrogen consumption of the whole vehicle is reduced, and the economic performance of the whole vehicle is effectively improved.
In the method for optimizing the fuzzy control strategy of the fuel cell vehicle based on the approximate model, in the process of optimizing the objective function, except the design space constraint of the design variable, the optimization process also ensures the driving requirement of the whole vehicle and the response speed of the fuel cell, and is represented by the following formula: pload(t)＝Pf(t)+Pb(t)，
Figure BDA0002548458240000031
Wherein, PfcscopeFor maximum dynamic rate of change of fuel cell output power, PbAnd outputting power for the power battery.
Compared with the prior art, the invention has the following beneficial effects:
1. the optimization method takes the relevant parameters of the membership functions in the fuzzy control strategy as variables to be optimized. And sampling the membership function variable in the fuzzy controller and the target value SOC of the power battery SOC by adopting an optimized Latin hypercube design to complete the test design. Sensitivity analysis is carried out on the variables to be optimized, the contribution rate of each variable to equivalent hydrogen consumption is calculated, the design variables with larger contribution rates are screened as optimization variables, and the precision and the optimization efficiency of the approximate model are improved.
2. The invention establishes an ellipsoid-based neural network complete vehicle simulation approximate model based on design variables. And sampling the screening design variables by adopting an optimized Latin hypercube design method, training the ellipsoidal-base neural network by utilizing sample data obtained by simulation, and establishing an approximate model. And randomly selecting a plurality of sampling points in a design space by adopting an optimized Latin hypercube design method, and evaluating the reliability of the approximate model by utilizing the simulated values and the observed values of the sampling points.
3. The optimization method adopts a multi-island genetic algorithm to optimize the related parameters of the membership function in the fuzzy control strategy. The output power of the fuel battery system after optimization is more stable than that before optimization, the equivalent hydrogen consumption is reduced, and the economic performance of the whole vehicle is effectively improved.
Drawings
FIG. 1 is a fuzzy distribution plot of power demand;
FIG. 2 is a fuzzy distribution map of SOC variations;
FIG. 3 is a fuzzy distribution plot of fuel cell system output power;
FIG. 4 is a schematic diagram of an optimal Latin hypercube design sampling point;
FIG. 5 is a first order Pareto pictorial view;
FIG. 6 is a second order Pareto illustration;
FIG. 7 is a schematic diagram of an ellipsoidal neural network model;
FIG. 8 is an approximation model error analysis;
FIG. 9 is a schematic flow diagram of a multi-island genetic algorithm;
FIG. 10 is a diagram of an iterative process of a multi-island genetic algorithm;
FIG. 11 is a graph comparing fuel cell system output power before and after optimization;
FIG. 12 is a graph comparing system efficiency of a fuel cell system before and after optimization;
fig. 13 is a fuel cell system efficiency map.
Detailed Description
The optimization method proposed by the present invention will be further illustrated and explained below with reference to the accompanying drawings.
A fuel cell vehicle energy fuzzy control strategy optimization method based on an approximate model comprises the following steps:
designing a fuzzy control energy management strategy for the whole fuel cell vehicle;
establishing an ellipsoid-based neural network whole vehicle simulation approximate model based on an optimized variable;
based on a multi-island genetic algorithm, relevant parameters of a membership function in a fuzzy control strategy are optimized.
1. And designing a whole vehicle energy fuzzy control management strategy of the fuel cell vehicle. The invention establishes a double-input single-output Mamdani fuzzy controller based on a fuzzy control strategy formulation principle and a power following control strategy control logic. Two input variables of the fuzzy controller are selected as the required power P of the whole vehicleloadAnd the SOC of the power battery is relative to the set value SOC*△ SOC, the fuzzy controller output variable is the fuel cell system reference output power Pf. The input-output variable's physical universe is set to: pload∈[-130,130]kW，△SOC∈[-0.55,0.45]，Pf∈[0,42.5]kW. Membership functions for each input and output quantity and their fuzzy distributions are shown in fig. 1-3.
And formulating a fuzzy control rule according to a fuzzy control strategy formulation principle and actual operation experience, wherein the following table is as follows:
Figure BDA0002548458240000041
because the formulation of the membership function and the fuzzy rule in the fuzzy controller mainly depends on engineering and practical experience, the fuzzy controller has larger contingency and subjectivity, so that the economic performance of the fuel cell is difficult to achieve global optimum, and the fuzzy controller has larger optimization space. The fuzzy rule is given by adopting conditional statements, parameterization is difficult to carry out, and the result after optimization is easy to cause to be not in accordance with the fuel cell automobile energy distribution principle, so that the membership function design parameters can be used as main optimization variables.
The whole vehicle model contains more nonlinear modules, so that the time for carrying out the economic simulation of the whole vehicle is longer, and the calculation and time cost required by optimization are overhigh due to more design variables. In order to improve the optimization efficiency, an approximate model can be established on the premise of ensuring the fitting precision, and the fuzzy control strategy of the fuel cell vehicle is optimized. Common approximate model methods mainly include polynomial models, response surface models, neural network models, and Kriging models. The neural network model has strong capability of approximating a complex nonlinear function and strong fault-tolerant capability, and can be used as a black box. The whole vehicle control is a highly nonlinear system, and an approximate model of an ellipsoid-based neural network can be selected for fitting.
2. And determining optimized design parameters, and establishing a complete vehicle simulation approximate model based on the ellipsoid-based neural network.
(1) And determining the optimized design parameters.
As known from the design of the fuzzy controller, the membership function can optimize the design parameters for 22 (x)1-x22) Meanwhile, the set value SOC of the power battery has a large influence on the economy of the whole vehicle, so that the fuzzy control strategy has 23 parameters to be optimized. Because the contribution rates of different parameters to the equivalent hydrogen consumption are different, the design parameters can be screened and optimized by calculating the contribution rates of the parameters, the number of the design parameters is reduced, and the precision and the optimization efficiency of the approximate model are improved. Firstly, the method takes equivalent hydrogen consumption as a response, adopts an optimized Latin hypercube design to sample 23 design parameters, and completes the experimental design. Fig. 4 shows a sample point profile for a 2-factor 9-level optimized latin hypercube design.
Next, sensitivity analysis was performed. According to the sample points and the response values thereof, a multiple quadratic regression model is established as follows:
Figure BDA0002548458240000051
β0、βi、βijis the model coefficient; normalizing input design variables to [ -1, 1]The model coefficient after least square fitting of the multiple quadratic regression model is SiWill SiConverted to a percentage contribution to the equivalent hydrogen consumption, reflecting the contribution of each input design variable to the response. And determining the final optimized design variable according to the influence result of the design variable on the equivalent hydrogen consumption. Fig. 5 and 6 are Pareto graphs of the results of the effect of each variable on the equivalent hydrogen consumption, which reflect the percentage contribution of all terms in the model to the response after sample fitting. Wherein fig. 5 represents the percentage of contribution of each variable to the equivalent hydrogen consumption, and fig. 6 represents the percentage of contribution of each variable to the equivalent hydrogen consumption including an interaction effect, wherein black represents a positive correlation and shaded represents a negative correlation. Comprehensively considering the response result and the characteristics of the power system components, and selecting X05、X07、X08、X10、X11、X12、X14、X17、X18、X19、X22And SOC*These 12 design parameters serve as optimization variables for the fuzzy control strategy.
(2) And establishing a complete vehicle simulation approximate model based on an ellipsoid-based neural network.
In order to construct an approximate model, an optimized Latin hypercube design method is adopted to sample 12 screened design parameters, and an ellipsoid-based neural network is trained by utilizing sample data obtained by simulation under an original finished automobile model to establish a finished automobile simulation approximate model. Fig. 7 is a schematic diagram of an ellipsoid-based neural network model, which is a feedforward neural network mainly composed of an input layer, a hidden layer, and an output layer.
The net input and net output of the hidden layer ellipsoidal neural unit can be represented by:
Figure BDA0002548458240000061
Figure BDA0002548458240000062
wherein z isj,iIs the initial semi-axial length of the ellipsoid unit function; w is ajb,iIs the semi-axis length of the ellipsoid-based neural unit relative to the input unit; w is ajc,iThe center of the ellipsoid-based neural unit relative to the input unit, and α is a slope adjusting parameter of the sigmoid curve, wherein the larger the value of the slope adjusting parameter is, the faster the speed of the output value of the ellipsoid-based function unit approaches zero is.
Taking the kth output of the ellipsoid-based neural network model as an example, assuming that there are n hidden layer units, the output can be expressed as:
Figure BDA0002548458240000063
wherein, wKjThe weight from hidden layer unit to output layer.
(3) And verifying the accuracy of the approximate model. And randomly selecting a plurality of sampling points in the design space, and evaluating the reliability of the approximate model by using the simulation values and the observation values of the approximate model of the sampling points. As shown in fig. 8, the observed values and the true values of the approximation model at each sampling point are more uniformly distributed near y ═ x, and the approximation accuracy of the approximation model based on the ellipsoid-based neural network meets the engineering requirements.
3. Based on a multi-island genetic algorithm, 12 design parameters of a membership function in a fuzzy control strategy are optimized.
The fuzzy control strategy is optimized by adopting a multi-island genetic algorithm, so that the fuel cell automobile realizes the best fuel economy, the flow chart of the optimization algorithm is shown in a flow chart 9, and the optimization problem is described by the following formula:
Figure BDA0002548458240000071
wherein the content of the first and second substances,
Figure BDA0002548458240000072
for the equivalent hydrogen consumption of the vehicle running, i.e. the objective function in the optimization process, the target value needs to be minimized for achieving the optimal economy of the vehicle running, X is an optimization design parameter, gi(X) is a constraint condition in the optimization process, and besides the design space constraint of design variables, the optimization process also ensures the driving required power of the whole vehicle and the response speed of the fuel cell, and is represented by the following formula:
Pload(t)＝Pf(t)+Pb(t) (6)
Figure BDA0002548458240000073
wherein, PfcscopeFor maximum dynamic rate of change of fuel cell output power, PbAnd outputting power for the power battery.
The optimization iterative process is shown in fig. 10, the equivalent hydrogen consumption shows a significant downward trend in the whole optimization process, and the algorithm has a good optimization function. And (3) redesigning a fuzzy control strategy according to the optimized variable value, and completing the optimization of the fuzzy control strategy of the fuel cell vehicle based on the approximate model, wherein the output power of the fuel cell system after the optimization is more stable than that before the optimization, the equivalent hydrogen consumption is reduced compared with that before the optimization, the equivalent hundred kilometer hydrogen consumption of the whole vehicle is reduced, and the economic performance of the whole vehicle is effectively improved. Fig. 11 is a graph of fuel cell system efficiency, fig. 12 is a graph showing a comparison of fuel cell system output power before and after optimization, and fig. 13 is a graph showing a comparison of fuel cell system efficiency before and after optimization.

Claims (2)

1. A fuel cell vehicle fuzzy control strategy optimization method based on an approximate model is characterized in that: the method comprises the following steps:
designing a fuel cell vehicle energy fuzzy control management strategy: the fuel cell vehicle Mamdani type fuzzy controller adopts a double-input single-output structural form, and two input variables of the fuzzy controller are selected as the required power P of the whole vehicleloadAnd the SOC of the power battery is relative to the set value SOC*△ SOC, the fuzzy controller output variable being the fuel cell systemTotal reference output power Pf(ii) a Defining fuzzy distribution of input and output quantities and formulating a fuzzy control rule;
establishing an ellipsoid-based neural network whole vehicle simulation approximate model related to an optimized variable:
(1) determining an optimal design variable: screening design variables of input and output quantity membership functions by calculating parameter contribution rates, and reducing the number of the design variables; firstly, sampling design variables by adopting an optimized Latin hypercube method to complete test design; secondly, sensitivity analysis is carried out, a multiple quadratic regression model is established according to the sample points and the corresponding simulation output results y,
Figure FDA0002548458230000011
wherein, β0、βi、βijCoefficients of the regression model; normalizing input design variables to [ -1, 1]Fitting the multiple quadratic regression model by using a least square method to obtain a model coefficient SiWill SiConverting into percentage of contribution rate to the equivalent hydrogen consumption, and reflecting the contribution of each design variable to the response-the equivalent hydrogen consumption; determining a fuzzy control strategy design variable which needs to be optimized finally according to the influence result of the design variable on the equivalent hydrogen consumption;
(2) establishing an overall vehicle simulation approximate model based on an ellipsoid-based neural network:
sampling the screened design variables to be optimized finally by adopting an optimized Latin hypercube design method, and training an ellipsoid-based neural network by utilizing the obtained sample data, thereby establishing a complete vehicle simulation approximate model about the optimized design variables;
optimizing a fuzzy control strategy by utilizing a multi-island genetic algorithm based on a complete vehicle simulation approximate model: optimizing a fuzzy control strategy by adopting a multi-island genetic algorithm to ensure that the fuel cell automobile realizes the optimal economy of the whole automobile; the optimization problem is described by the following formula:
Figure FDA0002548458230000012
wherein the content of the first and second substances,
Figure FDA0002548458230000013
for the equivalent hydrogen consumption of the vehicle, i.e. the objective function in the optimization process, to achieve optimum economy of the entire vehicle travel, the target value needs to be minimized, giAnd (X) redesigning the fuzzy control strategy according to the optimized design variable as a constraint condition in the optimization process, and completing the optimization of the fuel cell automobile fuzzy control strategy based on the approximate model.
2. The fuel cell vehicle fuzzy control strategy optimization method based on the approximate model as claimed in claim 1, characterized in that: in the process of optimizing the objective function, besides the design space constraint of the design variables, the optimization process should also ensure the driving requirement of the whole vehicle and the response speed of the fuel cell, which is expressed by the following formula: pload(t)＝Pf(t)+Pb(t)，
Figure FDA0002548458230000021
Wherein, PfcscopeFor maximum dynamic rate of change of fuel cell output power, PbAnd outputting power for the power battery.
CN202010568570.6A 2020-06-19 2020-06-19 Fuel cell vehicle fuzzy control strategy optimization method based on approximate model Active CN111695202B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010568570.6A CN111695202B (en) 2020-06-19 2020-06-19 Fuel cell vehicle fuzzy control strategy optimization method based on approximate model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010568570.6A CN111695202B (en) 2020-06-19 2020-06-19 Fuel cell vehicle fuzzy control strategy optimization method based on approximate model

Publications (2)

Publication Number Publication Date
CN111695202A true CN111695202A (en) 2020-09-22
CN111695202B CN111695202B (en) 2022-04-12

Family

ID=72482310

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010568570.6A Active CN111695202B (en) 2020-06-19 2020-06-19 Fuel cell vehicle fuzzy control strategy optimization method based on approximate model

Country Status (1)

Country Link
CN (1) CN111695202B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112455420A (en) * 2020-10-23 2021-03-09 西安交通大学 Hybrid power system energy control method based on fuzzy neural network
CN112531187A (en) * 2020-12-09 2021-03-19 奇瑞汽车股份有限公司 Fuel cell oxygen ratio control method, device and computer storage medium
CN113263960A (en) * 2021-06-28 2021-08-17 太原理工大学 Self-adaptive energy management method for hydrogen fuel cell automobile
CN113352946A (en) * 2021-07-05 2021-09-07 太原理工大学 Energy management method of fuel cell automobile power system
CN114488821A (en) * 2022-04-06 2022-05-13 国网浙江省电力有限公司电力科学研究院 Method and system for prediction control of interval economic model of fuel cell oxygen ratio
CN116278987A (en) * 2023-03-16 2023-06-23 佛山仙湖实验室 Hydrogen fuel cell automobile energy management method and system based on pigeon optimization algorithm

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050256631A1 (en) * 2004-05-14 2005-11-17 Cawthorne William R Method of determining engine output power in a hybrid electric vehicle
US20160137068A1 (en) * 2014-11-14 2016-05-19 Toyota Jidosha Kabushiki Kaisha Braking force control system, vehicle and method of controlling braking force
CN106064623A (en) * 2016-06-03 2016-11-02 北京理工大学 Motor vehicle driven by mixed power is gearshift control optimization method in braking procedure
CN108312870A (en) * 2018-02-02 2018-07-24 杭州电子科技大学 A kind of energy management method of hybrid vehicle hydrogen consumption and load variation

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050256631A1 (en) * 2004-05-14 2005-11-17 Cawthorne William R Method of determining engine output power in a hybrid electric vehicle
US20160137068A1 (en) * 2014-11-14 2016-05-19 Toyota Jidosha Kabushiki Kaisha Braking force control system, vehicle and method of controlling braking force
CN106064623A (en) * 2016-06-03 2016-11-02 北京理工大学 Motor vehicle driven by mixed power is gearshift control optimization method in braking procedure
CN108312870A (en) * 2018-02-02 2018-07-24 杭州电子科技大学 A kind of energy management method of hybrid vehicle hydrogen consumption and load variation

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
M. ISHAQBHATTI等: ""Recent development in copula and its applications to the energy, forestry and environmental sciences"", 《INTERNATIONAL JOURNAL OF HYDROGEN ENERGY》 *
卢云 等: ""某火箭炮托架结构的优化设计与研究"", 《机械制造与自动化》 *
吴茵: ""某纯电动MPV车内中高频噪声分析与声学包优化"", 《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅱ辑》 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112455420A (en) * 2020-10-23 2021-03-09 西安交通大学 Hybrid power system energy control method based on fuzzy neural network
CN112455420B (en) * 2020-10-23 2022-04-22 西安交通大学 Hybrid power system energy control method based on fuzzy neural network
CN112531187A (en) * 2020-12-09 2021-03-19 奇瑞汽车股份有限公司 Fuel cell oxygen ratio control method, device and computer storage medium
CN112531187B (en) * 2020-12-09 2022-05-03 奇瑞汽车股份有限公司 Fuel cell oxygen ratio control method, device and computer storage medium
CN113263960A (en) * 2021-06-28 2021-08-17 太原理工大学 Self-adaptive energy management method for hydrogen fuel cell automobile
CN113352946A (en) * 2021-07-05 2021-09-07 太原理工大学 Energy management method of fuel cell automobile power system
CN113352946B (en) * 2021-07-05 2022-10-04 太原理工大学 Energy management method of fuel cell automobile power system
CN114488821A (en) * 2022-04-06 2022-05-13 国网浙江省电力有限公司电力科学研究院 Method and system for prediction control of interval economic model of fuel cell oxygen ratio
CN114488821B (en) * 2022-04-06 2022-10-14 国网浙江省电力有限公司电力科学研究院 Method and system for predicting and controlling interval economic model of fuel cell oxygen passing ratio
CN116278987A (en) * 2023-03-16 2023-06-23 佛山仙湖实验室 Hydrogen fuel cell automobile energy management method and system based on pigeon optimization algorithm
CN116278987B (en) * 2023-03-16 2024-04-23 佛山仙湖实验室 Hydrogen fuel cell automobile energy management method and system based on pigeon optimization algorithm

Also Published As

Publication number Publication date
CN111695202B (en) 2022-04-12

Similar Documents

Publication Publication Date Title
CN111695202B (en) Fuel cell vehicle fuzzy control strategy optimization method based on approximate model
CN112776673B (en) Intelligent network fuel cell automobile real-time energy optimization management system
Pourkiaei et al. Modeling and experimental verification of a 25W fabricated PEM fuel cell by parametric and GMDH-type neural network
CN107451686B (en) Micro-grid energy optimization method of genetic algorithm considering random prediction error
CN113205207A (en) XGboost algorithm-based short-term power consumption load fluctuation prediction method and system
CN108306285B (en) Reliability evaluation method for active power distribution network containing distributed power supply based on TOPSIS method
Zhang et al. High-efficiency driving cycle generation using a Markov chain evolution algorithm
CN106786977B (en) Charging scheduling method of electric vehicle charging station
Xu et al. Research on double fuzzy control strategy for parallel hybrid electric vehicle based on GA and DP optimisation
CN110866636A (en) Microgrid planning method comprehensively considering electric vehicle charging station and distributed energy
CN111786420A (en) Power system economic dispatching method considering renewable energy hybrid power generation
CN110264012A (en) Renewable energy power combination prediction technique and system based on empirical mode decomposition
CN114418174A (en) Electric vehicle charging load prediction method
Lu et al. Online optimization of energy management strategy for FCV control parameters considering dual power source lifespan decay synergy
CN110570042B (en) Short-term electric vehicle charging load prediction method and system
CN117077960A (en) Day-ahead scheduling optimization method for regional comprehensive energy system
Huang et al. Smart energy management system based on reconfigurable AI chip and electrical vehicles
CN115995810A (en) Wind power prediction method and system considering weather fluctuation self-adaptive matching
CN116632823A (en) Short-term photovoltaic power prediction method based on power conversion model and multi-layer perceptron
CN116345555A (en) CNN-ISCA-LSTM model-based short-term photovoltaic power generation power prediction method
CN110956304A (en) Distributed photovoltaic power generation capacity short-term prediction method based on GA-RBM
CN113435663A (en) CNN-LSTM combined load prediction method considering electric vehicle charging load influence
Lin et al. AER adaptive control strategy via energy prediction for PHEV
CN111509782A (en) Probabilistic power flow analysis method considering charging load and photovoltaic output random characteristics
CN112701679B (en) Active power distribution network short-term reliability evaluation method based on model-data hybrid drive

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant