CN101333961A - Hydrogen gas natural gas mixed fuel engine optimizing method - Google Patents
Hydrogen gas natural gas mixed fuel engine optimizing method Download PDFInfo
- Publication number
- CN101333961A CN101333961A CNA2008101182088A CN200810118208A CN101333961A CN 101333961 A CN101333961 A CN 101333961A CN A2008101182088 A CNA2008101182088 A CN A2008101182088A CN 200810118208 A CN200810118208 A CN 200810118208A CN 101333961 A CN101333961 A CN 101333961A
- Authority
- CN
- China
- Prior art keywords
- excess air
- ignition advance
- hydrogen
- advance angle
- air coefficient
- 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
Links
- UFHFLCQGNIYNRP-UHFFFAOYSA-N Hydrogen Chemical compound [H][H] UFHFLCQGNIYNRP-UHFFFAOYSA-N 0.000 title claims abstract description 91
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 title claims abstract description 32
- 238000000034 method Methods 0.000 title claims abstract description 22
- 239000000446 fuel Substances 0.000 title claims abstract description 18
- 239000003345 natural gas Substances 0.000 title claims abstract description 18
- 239000001257 hydrogen Substances 0.000 claims abstract description 97
- 229910052739 hydrogen Inorganic materials 0.000 claims abstract description 97
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 42
- 230000002068 genetic effect Effects 0.000 claims abstract description 20
- 238000002474 experimental method Methods 0.000 claims abstract description 19
- 238000003062 neural network model Methods 0.000 claims abstract description 6
- 239000007789 gas Substances 0.000 claims description 34
- 238000005457 optimization Methods 0.000 claims description 18
- 210000002569 neuron Anatomy 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 15
- 230000014509 gene expression Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 11
- 238000013400 design of experiment Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 claims description 3
- 238000003050 experimental design method Methods 0.000 claims description 3
- 230000035772 mutation Effects 0.000 claims description 3
- 238000013401 experimental design Methods 0.000 abstract description 4
- 239000000203 mixture Substances 0.000 description 28
- 238000012360 testing method Methods 0.000 description 17
- 238000010586 diagram Methods 0.000 description 11
- 239000011435 rock Substances 0.000 description 7
- 238000007599 discharging Methods 0.000 description 6
- 238000002485 combustion reaction Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 239000004215 Carbon black (E152) Substances 0.000 description 3
- 230000007812 deficiency Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 229930195733 hydrocarbon Natural products 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 108090000623 proteins and genes Proteins 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 150000002430 hydrocarbons Chemical class 0.000 description 2
- 150000002431 hydrogen Chemical class 0.000 description 2
- -1 hydrogen alkane Chemical class 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 230000000576 supplementary effect Effects 0.000 description 2
- GQPLMRYTRLFLPF-UHFFFAOYSA-N Nitrous Oxide Chemical class [O-][N+]#N GQPLMRYTRLFLPF-UHFFFAOYSA-N 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000010429 evolutionary process Effects 0.000 description 1
- 238000013100 final test Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 239000010705 motor oil Substances 0.000 description 1
- 238000010791 quenching Methods 0.000 description 1
- 230000000171 quenching effect Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Images
Classifications
-
- 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
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/30—Use of alternative fuels, e.g. biofuels
Landscapes
- Combined Controls Of Internal Combustion Engines (AREA)
- Electrical Control Of Air Or Fuel Supplied To Internal-Combustion Engine (AREA)
- Feedback Control In General (AREA)
Abstract
Disclosed is an optimizing method for a hydrogen and natural gas mixed fuel engine, which belongs to the hydrogen and natural gas mixed fuel engine technical field. The invention aims to effectively discover three control variables, including the optimum hydrogen mixing ratio, ignition advance angle and the excess air factor under each operating condition and obtain the optimum comprehensive performance of the engine. The rotation speed and the torque are used for determining the operating condition point; the three-dimension state space of the hydrogen mixing ratio, the ignition advance angle and the excess air factor is determined under the operating condition point; based on the method of experimental design, the most representative point is selected for the experiment; the data is acquired after the experiment; the points in the data are used for establishing the engine error backward neural network model; namely, the relation of the control variables of the hydrogen mixing ratio, ignition advance angle and the excess air factor, and the economical exhaust is established; the model data is exported and the genetic algorithm is adopted to optimize the engine control parameters; finally, the control variables of optimum hydrogen mixing ratio, the ignition advance angle and the excess air factor of the engine are acquired under the operating condition so that the optimum engine performance is obtained.
Description
Technical field
The present invention relates to the optimization method of hydrogen gas natural gas mixed fuel engine, belong to natural gas-hydrogen mixture engine fuel technical field.
Background technique
Natural gas-hydrogen mixture fuel is called for short HCNG, has another name called Hythane (hydrogen alkane), be hydrogen is mixed by a certain percentage with rock gas and obtain substitute gaseous fuel.It combines hydrogen and rock gas advantage separately: one, the velocity of combustion height of hydrogen, approximately be 8 times of rock gas, in rock gas, mix hydrogen and can improve the velocity of combustion of mixed gas, igniting can more close engine tope center, reduce the compression negative work, improve burning constant volume degree, can improve the thermal efficiency; Two, the flammable mixing boundary of hydrogen is wide, and the lean combustion limit reaches 0.068 (fuel air ratio relatively); Quenching distance has only 30% of rock gas.Therefore, small quantity of hydrogen can be widened the flammable mixed proportion of mixed gas, can realize lean combustion; Three, the volume calorific value of rock gas is higher, has higher volumetric heat than pure hydrogen.
Mix hydrogen than referring to the shared volume ratio of hydrogen in the HCNG fuel.Emission performance is comprehensive emission performance, comprises CO, CH
4, NOX discharging; Economy is meant the gas consumption rate at fuel, because the quality calorific value of hydrogen is 2.4 times of the quality of natural gas calorific value, the specific fuel consumption of fuel combination is inevitable behind the natural gas-hydrogen mixture reduces along with the increase of mixing the hydrogen ratio, in order to contrast the capacity usage ratio under the suitable situation of mixed gas fuel value, the hydrogen that consumes in the mixed gas is converted into quality of natural gas with calorific value, thereby obtains the equivalent proportion gas consumption.
About the combustion characteristic of HCNG fuel on motor, also have a lot of scholars to do research both at home and abroad: people such as Munshi have carried out 20% to 32% and have mixed hydrogen than experiment (S.R.Munshi on a turbosupercharging lean burn natural gas engine, C.Nedelcu, J.Harris et al.Hydrogen Blended Natural Gas Operation of a Heavy Duty TurbochargedLean Burn Spark Ignition Engine.SAE papers 2004-01-2956), the result shows, mix hydrogen than being that 20% to 30% mixed air is under the prerequisite that does not influence power character and efficient, can at utmost reduce the NOx discharging, can keep hydrocarbon emission simultaneously and be lower than the raw natural gas hydrocarbon emission.In addition, mixing needs behind the hydrogen to strengthen excess air coefficient and reduce ignition advance angle, so just can make the motor combination property reach optimum.Sierens etc. carry out difference and mix hydrogen than experiment (R.Sierens on the motor of a general V8, E.Rosseel.Variable Composition Hydrogen/Natural Gas Mixtures forIncreased Engine Efficiency and Decreased Emissions.Journal of Engineering for GasTurbines and Power.January 2000, Vol.122/135.), the result shows: mix the hydrogen ratio less than 20% o'clock, must be in order to reduce the hydrocarbon emission excess air coefficient less than 1.3, but must be in order to reduce the discharged nitrous oxides excess air coefficient greater than 1.5, mix that there is a difficult problem in the optimization of excess air coefficient behind the hydrogen.
Above-mentioned research mainly is after mixing hydrogen engine performance to be influenced research, and optimize only at ignition advance angle or excess air coefficient, make the motor combination property reach optimum, must be to mixing the hydrogen ratio, ignition advance angle and excess air coefficient are optimized simultaneously.
Because three controlled variable are mixed hydrogen ratio, excess air coefficient and ignition advance angle to the motor Economy, there is the relation of coupling in emission performance, therefore be the multiobject optimization problem of ternary, if exist experimental amount big according to traditional experimental technique, and can not obtain the problem of optimum Control Parameter, therefore need to seek new method and solve this difficult problem.
Summary of the invention
The objective of the invention is to propose a kind of optimization method of hydrogen gas natural gas mixed fuel engine, this method can be optimized mixing hydrogen ratio, ignition advance angle and excess air coefficient on the basis of saving a large amount of experiments simultaneously, can increase substantially the optimization precision simultaneously.
Technological scheme of the present invention is as follows:
The optimization method of hydrogen gas natural gas mixed fuel engine, carry out as follows:
1) design of experiment:
Selecting arbitrary rotating speed of motor and moment of torsion operating point, is controlled variable to mix hydrogen ratio, ignition advance angle and excess air coefficient, and mixing hydrogen volume percentage is 0～40%, and ignition advance angle is 20 °～40 °, and excess air coefficient is 1.1～1.6; Adopt the orthogonalization experimental design method, at described ignition advance angle, excess air coefficient with mix the combination that hydrogen is selected 9 groups of controlled variable in than regulation range, do experiment with these 9 groups of controlled variable respectively, obtain every group of gas consumption rate, NOx, CH that experiment is corresponding
4With the CO target variable; Adopt the extreme difference methods analyst to mix hydrogen ratio, ignition advance angle, excess air coefficient to gas consumption rate, NOx, CH
4With CO influence power size, obtain mixing hydrogen than 10%～40%, 20 °～30 ° of ignition advance angles and excess air coefficient 1.3～1.6 are the zone of optimal control parameter, more careful design of experiment is carried out in this zone, finally obtain the combination of 30 groups of controlled variable, adopt these 30 groups of controlled variable to do experiment, obtain laboratory data;
2) set up error anti-pass neural network model:
A. data pretreatment: ignition advance angle, excess air coefficient is mixed hydrogen than between the controlled variable linear mapping to 0 to 1, and the mapping formula is:
In the formula: Xn is the value of mixing hydrogen ratio, ignition advance angle and excess air coefficient after the normalization, and Xi is the actual value of mixing hydrogen ratio, ignition advance angle and excess air coefficient; Xmax and Xmin are respectively the maximum value and the minimum value of mixing hydrogen ratio, ignition advance angle and excess air coefficient;
B. set up error anti-pass neural network model, parameter-definition is: the node number of input layer selects 3, corresponding ignition advance angle, excess air coefficient and mix the hydrogen ratio; Hidden layer is 2 layers, and first hidden neuron selects 10, and second layer neuron selects 5; Output layer node number selects 4, corresponding gas consumption rate, NOx, CH
4And CO; The training function is selected the Levenberg-Marquardt algorithm; Learning rate selects 0.01;
C. neural network training: the laboratory data that obtains in the design of experiment is used for training error anti-pass neuron network, restrains up to error anti-pass neuron network;
3) adopt the genetic algorithm optimization Control Parameter:
A. genetic algorithm parameter setting: it is 40 that genetic algorithm program population size is set, and maximum evolutionary generation is 80, and the generation gap in different generations is 0.8;
B, fitness function setting: determine NOx, CH
4, CO emissions object variable is identical with the weight of gas consumption rate Economy target variable, sets up the fitness function that shows individual superiority-inferiority:
Objectfunction=(CH
4(model value)/CH
4(reference value))
2+ (CO (model value)/CO (reference value))
2+ (NOx (model value)/NOx (reference value))
2+ (BSFC (model value)/BSFC (reference value))
2
In the formula: objectfunction represents fitness function size, CH
4(model value), CO (model value), NOx (model value) and BSFC (model value) expression are mixed the hydrogen ratio for corresponding one group, ignition advance angle and excess air coefficient model calculate CH
4, CO, NOx and gas consumption rate; CH4 (reference value), CO (reference value), NOx (reference value) and BSFC (reference value) expression CH
4, CO, NOx and gas consumption rate correspondence reference value, be respectively 1500,1000,2500,270;
C, selection cross and variation operator are provided with: selection algorithm is selected the roulette selection algorithm; Crossover algorithm is selected even crossover algorithm, and crossover probability selects 0.7; The variation algorithm is selected basic mutation operator, and the variation probability selects 0.01;
D, by selecting, intersect and variation, obtain population of future generation, turn back to step b again, circulation repeatedly, up to evolutionary generation reach that step a is provided with 80 till, the output optimal result.
The present invention is experimental design, neuron network and genetic algorithm are used for the optimization of hydrogen-blended natural gas engine Control Parameter, successfully solved the coupled problem of mixing hydrogen ratio, ignition advance angle and three controlled variable of excess air coefficient, Economy and emission performance are optimized simultaneously, and the combination property of motor is improved.
Description of drawings
Fig. 1 is the entire block diagram of hydrogen-blended natural gas engine optimization method.
Fig. 2 is the experimental system hardware block diagram.
Fig. 3 is the genetic algorithm program block diagram.
Among Fig. 2: the 1-natural gas cylinder; The 2-hydrogen cylinder; The 3-natural gas flowmeter; 4-hydrogen flowing quantity controller; The 5-hydrogen flowmeter; 6-mixed stable voltage jar; 7-mixes the hydrogen system control module; The 8-motor; The 9-ignition key; 10-engine control module (ECU); A-hydrogen flowing quantity signal; B-gas discharge signal;-engine rotational speed signal;-engine oil door position signal; The e-fire signal; F-hydrogen flowing quantity control signal; G-hydrogen electromagnetic valve signal; H-rock gas electromagnetic valve signal.
Embodiment
Below in conjunction with drawings and Examples principle of the present invention, working procedure are further described.Selecting operating point among the embodiment is rotating speed 1200r/min, moment of torsion 200N.M, and the ignition advance angle excursion is the 20-40 degree, and the excess air coefficient regulation range is 1.1-1.6, and mixing hydrogen is 0-40% than regulation range.The optimal control method overall flow is as shown in Figure 1:
1) design of experiment
Adopting the design of experiment main purpose is to obtain most important information with minimum experiment.From mixing hydrogen ratio, ignition advance angle and excess air coefficient regulation range and each regulated quantity,, need do 330 experiments if all do experiment at all experimental points.This is the experimental amount that a rotating speed moment of torsion operating point needs, and optimizes a plurality of rotating speed moment of torsion operating points if desired, and experimental amount also can become multiple to increase.And the experimental design method among employing the present invention, only needs are tested for 30 times, just can obtain the data of needs, have significantly reduced the experimentation work amount.
A) orthogonalization design of experiment determines to mix hydrogen ratio, ignition advance angle, excess air coefficient to Economy and emission performance influence power size.
The essence of orthogonal experimental method is to choose a part of representational combination and tests from whole composite tests.Its characteristics are that the test point is representative strong, and test data is balanced disperses, and test data is neatly comparable.Orthogonal experiment can obtain test effect preferably by less test number (TN).Because three Control Parameter are arranged, option table 1 orthogonal test table of the present invention is selected 9 points in the Control Parameter space.These 9 design points are done experiment, obtain test result.The detail parameters and the test result thereof of 9 design points see Table 1.
Table 1 orthogonalization experimental design point and test result thereof
| Tested number | Ignition advance angle | Mix the hydrogen ratio | Excess air coefficient | The gas consumption rate | CO | CH 4 | NO x |
| 1 | 40 | 0 | 1.1 | 296.27 | 1450 | 1805 | 3813 |
| 2 | 30 | 0 | 1.3 | 272.08 | 456 | 1050 | 3330 |
| 3 | 20 | 0 | 1.6 | 279.64 | 390 | 1394 | 56 |
| 4 | 20 | 0.2 | 1.1 | 280.95 | 562 | 872 | 3152 |
| 5 | 30 | 0.2 | 1.6 | 258.49 | 375 | 1549 | 278 |
| 6 | 40 | 0.2 | 1.3 | 282.44 | 403 | 861 | 5009 |
| 7 | 40 | 0.4 | 1.1 | 308.94 | 415 | 807 | 5000 |
| 8 | 30 | 0.4 | 1.3 | 271.84 | 291 | 481 | 4890 |
| 9 | 20 | 0.4 | 1.6 | 256.86 | 365 | 877 | 252 |
B) test result is carried out range analysis
The difference that maximum value in one group of data deducts the minimum value gained is called the extreme difference of these group data, the extreme difference reflection be excursion or amplitude of variation of these group data.The controlled variable that extreme difference is big illustrates that this controlled variable is big to the influence of experimental result.Utilize range analysis to investigate comparatively intuitively and mix hydrogen ratio, ignition advance angle and excess air coefficient change influence power size in the test, thereby can point the direction for further experimental design to gas consumption rate and each effulent.Table 2,3,4,5 is respectively NOx, COBSFC range analysis.
Table 2 NOx range analysis
| Position level 1 | Position level 2 | Position level 3 | Extreme difference | |
| Ignition advance angle | 3460 | 8498 | 13822 | 10362 |
| Mix the hydrogen ratio | 7199 | 8439 | 10142 | 2943 |
| Excess air coefficient | 11965 | 13229 | 586 | 12643 |
Attached: the position level 1 expression ignition advance angle of ignition advance angle correspondence is three NOx dischargings sum, i.e. 56+3152+252=3460 in the table 1 of 20 correspondences when spending in the table 2; The position level 2 expression ignition advance angles of ignition advance angle correspondence are three NOx discharging sums of 30 correspondences when spending; The position level 3 expression ignition advance angles of ignition advance angle correspondence are three NOx discharging sums of 40 correspondences when spending, and the extreme difference of ignition advance angle correspondence deducts minimum value for the maximum value in the position level 1,2,3, i.e. 13822-3460=10362.In like manner, mix hydrogen than corresponding position level 1,2,3 expressions mix hydrogen than be 0,20%, three NOx of 40% o'clock correspondence discharge sums, position level 1,2, the 3 expression excess air coefficients of excess air coefficient correspondence are three NOx discharging sums of 1.1,1.3,1.6 o'clock correspondences, its separately extreme difference deduct minimum value for the maximum value in the position level of its correspondence.Table 3,4,5 can the rest may be inferred.
Table 3 CO range analysis
| Position level 1 | Position level 2 | Position level 3 | Extreme difference | |
| Ignition advance angle | 1317 | 1122 | 2268 | 1146 |
| Mix the hydrogen ratio | 2296 | 1340 | 1071 | 1225 |
| Excess air coefficient | 2427 | 1150 | 1130 | 1297 |
Table 4 CH
4Range analysis
| Position level 1 | Position level 2 | Position level 3 | Extreme difference | |
| Ignition advance angle | 3143 | 3080 | 3473 | 393 |
| Mix the hydrogen ratio | 4249 | 3282 | 2165 | 2084 |
| Excess air coefficient | 3484 | 2392 | 3820 | 1428 |
Table 5 BSFC range analysis
| Position level 1 | Position level 2 | Position level 3 | Extreme difference | |
| Ignition advance angle | 817.45 | 802.41 | 887.65 | 85.24 |
| Mix the hydrogen ratio | 847.99 | 821.88 | 837.64 | 26.11 |
| Excess air coefficient | 886.16 | 826.36 | 794.99 | 91.47 |
Get by the big I of extreme difference relatively, mix hydrogen and be respectively: excess air coefficient＞ignition advance angle＞mix hydrogen ratio than, ignition advance angle and excess air coefficient influence power size to NOx; Mixing hydrogen is respectively than, ignition advance angle and the excess air coefficient influence power size to CO: excess air coefficient＞mix hydrogen ratio＞ignition advance angle; Mix hydrogen ratio, ignition advance angle and excess air coefficient to CH
4The influence power size be respectively: mix hydrogen ratio＞excess air coefficient＞ignition advance angle; Mixing hydrogen is respectively than, ignition advance angle and the excess air coefficient influence power size to BSFC: excess air coefficient＞ignition advance angle＞mix hydrogen ratio.Because ignition advance angle is increased to 40 degree, NOx, CO, CH from 20 degree
4, the position level of ignition advance angle correspondence increases gradually among the BSFC, illustrates that the optimum igniting advance angle is between 20-30; Increase NOx, CO, CH with excess air coefficient
4, the position level of excess air coefficient correspondence increases gradually among the BSFC, illustrates that best excess air coefficient is between 1.3-1.6; After mixing hydrogen, CO, CH
4, BSFC reduces, and illustrates that the best mixes hydrogen than between 10%-40%.
Finally obtain mixing hydrogen than 10%～40%, 20 °～30 ° of ignition advance angles and excess air coefficient 1.3～1.6 are the zone of optimal control parameter.
C) be foundation with the influence power size, the supplementary test design
According to the range analysis result, mixing hydrogen than 10%～40%, 21 test points that 20 °～30 ° of ignition advance angles and excess air coefficient 1.3～1.6 regional supplements are outer make the test sum reach 30, and final test point is as shown in table 6.Do experiment at these points, obtain to mix hydrogen ratio, ignition advance angle and excess air coefficient and gas consumption rate, NOx, CH
4, CO corresponding relation.
| Tested number | Mix the hydrogen ratio | Ignition advance angle | Excess air coefficient |
| 1 | 0 | 40 | 1.1 |
| 2 | 0 | 30 | 1.3 |
| 3 | 0 | 20 | 1.6 |
| 4 | 0.2 | 20 | 1.1 |
| 5 | 0.2 | 30 | 1.6 |
| 6 | 0.2 | 40 | 1.3 |
| 7 | 0.4 | 40 | 1.1 |
| 8 | 0.4 | 30 | 1.3 |
| 9 | 0.4 | 20 | 1.6 |
| 10 | 0.35 | 40 | 1.1 |
| 11 | 0.35 | 30 | 1.3 |
| 12 | 0.35 | 20 | 1.6 |
| 13 | 0.3 | 40 | 1.1 |
| 14 | 0.3 | 30 | 1.3 |
| 15 | 0.3 | 20 | 1.6 |
| 16 | 0.15 | 40 | 1.5 |
| 17 | 0.2 | 30 | 1.5 |
| 18 | 0.25 | 20 | 1.5 |
| 19 | 0.3 | 20 | 1.5 |
| 20 | 0.15 | 20 | 1.45 |
| 21 | 0.2 | 25 | 1.45 |
| 22 | 0.25 | 30 | 1.45 |
| 23 | 0.3 | 35 | 1.45 |
| 24 | 0.1 | 20 | 1.4 |
| 25 | 0.2 | 25 | 1.4 |
| 26 | 0.3 | 30 | 1.4 |
| 27 | 0.35 | 35 | 1.4 |
| 28 | 0.2 | 25 | 1.2 |
| 29 | 0.25 | 20 | 1.2 |
| 30 | 0.3 | 30 | 1.2 |
2) set up error anti-pass neural model
The purpose of setting up model is to obtain to mix hydrogen ratio, ignition advance angle and excess air coefficient input controlled variable to gas consumption rate, NOx, CH
4, CO export target variable mapping relations, serve as three kinds of gas (CH that output research Abgasgesetz comprises to mix hydrogen than, ignition advance angle and excess air coefficient
4, CO and NOx) and the individual features of the Economy of motor.
A. data pretreatment: because the excursion separately of input variable is very big, mapping model may be by the match modeling, need this moment to calculate each input quantity and determine that correlation coefficient then can cause the weight of each input variable inconsistent, thereby influence the factor that influences of the less input quantity of value; If dependence neural net model establishing, then excessive input quantity might make neuron saturated.Data normalization can be made respectively import component status of equal importance is arranged, prevent that the big output component absolute error of numerical value is big, the output component absolute error that numerical value is little is little, thereby helps the effect weights adjusted according to total error.Ignition advance angle, excess air coefficient is mixed hydrogen than between the controlled variable linear mapping to 0 to 1, and the mapping formula is:
In the formula: Xn is the value of mixing hydrogen ratio, ignition advance angle and excess air coefficient after the normalization, and Xi is the actual value of mixing hydrogen ratio, ignition advance angle and excess air coefficient; Xmax and Xmin are respectively the maximum value and the minimum value of mixing hydrogen ratio, ignition advance angle and excess air coefficient;
B. set up error anti-pass neural network model: set up error anti-pass neuron network and need prevent match deficiency and overfitting.If matched curve is the variation tendency of tracking response characteristic accurately, this situation is exactly the match deficiency.Opposite, if interference of noise is received in matched curve, thereby promptly constantly press close to noise complicated response characteristic, this has just caused overfitting.
The not enough main influence factor of overfitting and match is to introduce the parameter of model.Parameter is excessive, and the matched curve complexity then may cause overfitting; On the contrary, parameter is very few, and matched curve is simple relatively, thus then correctly the tracking response characteristic cause the match deficiency.
Through adjusting repeatedly, Select Error anti-pass neural network parameter is: the node number of input layer selects 3, corresponding controlled variable number; Hidden layer is 2 layers, and first hidden neuron selects 10, and second layer neuron selects 5; Output layer node number selects 4, corresponding target variable number; The training function is selected the Levenberg-Marquardt algorithm; Learning rate selects 0.01;
C. neural network training: the laboratory data that obtains in the design of experiment is used for training error anti-pass neuron network, and up to the convergence of error anti-pass neuron network, training is finished, and neuron network successfully constructs.
3) adopt the genetic algorithm optimization Control Parameter:
Genetic algorithm is the computation model of the biological evolution process of Darwinian heredity selection of simulation and natural selection, is a kind of method by simulating nature evolutionary process search optimal solution.Genetic algorithm is the searching algorithm with robustness that a class can be used for complex system optimization, compare with traditional optimized Algorithm, mainly contain following characteristics: genetic algorithm directly with objective function as search information, and traditional optimized Algorithm not only needs target function value, but also needs some other supplementarys such as derivative value of objective function could determine the direction of search.This characteristics make genetic algorithm only need that performance is carried out in the output of system and pass judgment on, have nothing to do with system complexity, and be a kind of method of black box.And genetic algorithm is convenient to handle the multi-dimensional optimization problem, therefore relatively more suitable control parameters of engine optimization problem.Fig. 3 is the genetic algorithm program block diagram, solid arrow representation program flow process wherein, and dotted arrow is represented the explanation to this flow process.
A. genetic algorithm parameter setting: in this step corresponding diagram 3 (1).It is 40 that genetic algorithm program population size is set, and maximum evolutionary generation is 80, and the generation gap in different generations is 0.8, and these parameters are provided with mainly and choose by experience, and principle is a fast convergence rate, optimizes the precision height simultaneously, and these three parameters are by the repetition test gained;
B. the genetic algorithm fitness function is provided with: in this step corresponding diagram 3 (2).Use fitness function to weigh each individuality in the colony can reach or approach to find optimal solution in computation optimization good degree in the genetic algorithm.It is bigger that follow-on probability is arrived in the higher individual inheritance of fitness.Consider that emission performance and Economy are of equal importance to motor, NOx, CH are set
4, CO emissions object variable and gas consumption rate Economy target variable weight identical all be 1, set up the fitness function that shows individual superiority-inferiority:
Objectfunction=(CH
4(model value)/CH
4(reference value))
2+ (CO (model value)/CO (reference value))
2+ (NOx (model value)/NOx (reference value))
2+ (BSFC (model value)/BSFC (reference value))
2
In the formula: CH
4(model value), CO (model value), NOx (model value) and BSFC (model value) expression are mixed the hydrogen ratio for corresponding one group, ignition advance angle and excess air coefficient model calculate CH
4, CO, NOx and gas consumption rate; CH
4(reference value), CO (reference value), NOx (reference value) and BSFC (reference value) expression CH
4, CO, NOx and gas consumption rate correspondence reference value, be respectively 1500,1000,2500,270;
C. judge fitness: in this step corresponding diagram 3 (3).Calculate the output of mixing hydrogen ratio, ignition advance angle and excess air coefficient for input according to the model of having set up, and ordering, as the criterion of determining fitness.If obtain the fitness peak, then directly output; If no, then proceed to select the cross and variation algorithm;
D. selection algorithm setting: in this step corresponding diagram 3 (4).It is to select some better individualities from current population that this step corresponding diagram is selected the operator effect, and it is copied in the population of future generation.Selection algorithm is selected the roulette selection algorithm, and the selection probability of each individuality and its fitness value are proportional in this algorithm.
E. crossover algorithm setting: (5) in this step corresponding diagram 3.Crossover algorithm makes that on the one hand the characteristic of defect individual can keep to a certain extent in original population, also makes algorithm can explore new gene space on the other hand, thereby makes the individuality in the new colony have diversity.Crossover algorithm is selected even crossover algorithm, and crossover probability selects 0.7.Evenly intersection exchanges each of two former generation's genes of individuals strings according to probability.
F. the algorithm setting makes a variation: (6) in this step corresponding diagram 3.The substance of variation algorithm is that the genic value on some gene location of the individuality string in the colony is changed.The variation algorithm is selected basic mutation operator, and the variation probability selects 0.01.
G. by selecting, intersect and variation, obtain population of future generation, turn back to step (2) again, circulation repeatedly, up to evolutionary generation reach that step (1) is provided with 80 till;
H. optimal result: through above optimization, obtain the best Control Parameter of fitness, this routine optimal control parameter scope is 22 °～25 °, 1.3～1.35 and 30%.
Claims (1)
1. the optimization method of hydrogen gas natural gas mixed fuel engine is characterized in that this method carries out as follows:
1) design of experiment:
Selecting arbitrary rotating speed of motor and moment of torsion operating point, is controlled variable to mix hydrogen ratio, ignition advance angle and excess air coefficient, and mixing hydrogen volume percentage is 0～40%, and ignition advance angle is 20 °～40 °, and excess air coefficient is 1.1～1.6; Adopt the orthogonalization experimental design method, at described ignition advance angle, excess air coefficient with mix the combination that hydrogen is selected 9 groups of controlled variable in than regulation range, do experiment with these 9 groups of controlled variable respectively, obtain every group of gas consumption rate, NOx, CH that experiment is corresponding
4With the CO target variable; Adopt the extreme difference methods analyst to mix hydrogen ratio, ignition advance angle, excess air coefficient to gas consumption rate, NOx, CH
4With CO influence power size, obtain mixing under the described operating point hydrogen than 10%～40%, 20 °～30 ° of ignition advance angles and excess air coefficient 1.3～1.6 be the zone of optimal control parameter, in this zone, select 21 groups of controlled variable again, finally obtain the combination of 30 groups of controlled variable, adopt these 30 groups of controlled variable to do experiment, obtain laboratory data;
2) set up error anti-pass neural network model:
A. data pretreatment: the ignition advance angle in the laboratory data, excess air coefficient is mixed hydrogen than between the controlled variable linear mapping to 0 to 1, and the mapping formula is:
In the formula: Xn is the value of mixing hydrogen ratio, ignition advance angle or excess air coefficient after the normalization, and Xi is the actual value of mixing hydrogen ratio, ignition advance angle or excess air coefficient; Xmax and Xmin are respectively the maximum value and the minimum value of mixing hydrogen ratio, ignition advance angle or excess air coefficient;
B. set up error anti-pass neural network model, parameter-definition is: the node number of input layer selects 3, corresponding ignition advance angle, excess air coefficient and mix the hydrogen ratio; Hidden layer is 2 layers, and first hidden neuron selects 10, and second layer neuron selects 5; Output layer node number selects 4, corresponding gas consumption rate, NOx, CH
4And CO; The training function is selected the Levenberg-Marquardt algorithm; Learning rate selects 0.01, and " Levenberg-Marquardt " is a kind of training algorithm;
C. neural network training: the data of handling well among the step a are used for the error anti-pass neuron network that training step b is provided with, restrain up to error anti-pass neuron network;
3) adopt the genetic algorithm optimization Control Parameter:
A. genetic algorithm parameter setting: it is 40 that genetic algorithm program population size is set, and maximum evolutionary generation is 80, and the generation gap in different generations is 0.8;
B, fitness function setting: determine NOx, CH
4, CO is identical with the weight of gas consumption rate target variable, sets up the fitness function that shows individual superiority-inferiority:
Objectfunction=(CH
4(model value)/CH
4(reference value))
2+ (CO (model value)/CO (reference value))
2+ (NOx (model value)/NOx (reference value))
2+ (BSFC (model value)/BSFC (reference value))
2
In the formula: objectfunction represents fitness function size, CH
4(model value), CO (model value), NOx (model value) and BSFC (model value) expression are mixed the hydrogen ratio for corresponding one group, ignition advance angle and excess air coefficient model calculate CH
4, CO, NOx and gas consumption rate; CH
4(reference value), CO (reference value), NOx (reference value) and BSFC (reference value) expression CH
4, CO, NOx and gas consumption rate correspondence reference value, be respectively 1500,1000,2500,270;
C, selection cross and variation operator are provided with: selection algorithm is the roulette selection algorithm; Crossover algorithm is even crossover algorithm, and crossover probability selects 0.7; The variation algorithm is basic mutation operator, and the variation probability selects 0.01;
D, by selecting, intersect and variation, obtain population of future generation, turn back to step b again, circulation repeatedly, up to evolutionary generation reach that step a is provided with 80 till; The output optimal result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008101182088A CN101333961B (en) | 2008-08-07 | 2008-08-07 | Hydrogen gas natural gas mixed fuel engine optimizing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008101182088A CN101333961B (en) | 2008-08-07 | 2008-08-07 | Hydrogen gas natural gas mixed fuel engine optimizing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101333961A true CN101333961A (en) | 2008-12-31 |
| CN101333961B CN101333961B (en) | 2011-12-28 |
Family
ID=40196783
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008101182088A Active CN101333961B (en) | 2008-08-07 | 2008-08-07 | Hydrogen gas natural gas mixed fuel engine optimizing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101333961B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102337979A (en) * | 2011-08-11 | 2012-02-01 | 浙江大学 | Automatic calibration parameter optimization method of engine based on genetic algorithm |
| CN102626557A (en) * | 2012-04-13 | 2012-08-08 | 长春工业大学 | Molecular distillation process parameter optimizing method based on GA-BP (Genetic Algorithm-Back Propagation) algorithm |
| CN102717800A (en) * | 2012-06-13 | 2012-10-10 | 吉林大学 | Vehicle economy cruise control method based on engine universal characteristic |
| CN104598654A (en) * | 2014-10-07 | 2015-05-06 | 芜湖扬宇机电技术开发有限公司 | Ignition advance angle prediction system and method thereof |
| CN105020032A (en) * | 2015-07-31 | 2015-11-04 | 毛志明 | Novel self-adaptation gas fuel control system |
| CN106224088A (en) * | 2016-08-23 | 2016-12-14 | 石家庄新华能源环保科技股份有限公司 | A kind of method utilizing High Pressure Hydrogen fuel combination to drive automobile and dynamical system |
| CN108763681A (en) * | 2018-05-16 | 2018-11-06 | 华北水利水电大学 | Hydrogen engine fault diagnosis system and method based on FOA-GRNN blending algorithms |
| CN108757264A (en) * | 2018-04-26 | 2018-11-06 | 杭州电子科技大学 | A method of obtaining coke-oven gas engine Optimum spark advance angle |
| CN110080896A (en) * | 2019-04-24 | 2019-08-02 | 河南省图天新能源科技有限公司 | A kind of methane fuelled engine air/fuel ratio control method based on genetic algorithm |
| CN112257355A (en) * | 2020-10-30 | 2021-01-22 | 东南大学 | Modeling method for medium-low pressure gas distribution pipe network of hydrogen-doped natural gas |
| CN113337321A (en) * | 2021-05-17 | 2021-09-03 | 西安交通大学 | Use of natural gas as hydrogen flame retardant and fuel suitable for oxyhydrogen burner |
| US11174800B2 (en) | 2019-09-24 | 2021-11-16 | Caterpillar Inc. | Transient controller and method of operating gas engine |
| CN114483341A (en) * | 2022-01-12 | 2022-05-13 | 清华大学 | Combustion control method in hydrogen-doped natural gas engine |
| CN114961942A (en) * | 2022-06-14 | 2022-08-30 | 潍柴动力股份有限公司 | Method and device for reducing methane emission |
| CN115324698A (en) * | 2022-08-09 | 2022-11-11 | 上海交通大学 | Diesel engine tail gas analysis sensing method based on neural network |
| CN116006374A (en) * | 2023-02-20 | 2023-04-25 | 潍柴动力股份有限公司 | Fuel hydrogen-adding engine control method and device and vehicle |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11614040B1 (en) * | 2021-12-16 | 2023-03-28 | Caterpillar Inc. | Cost based substitution ratio control in hydrogen-capable internal combustion engine system |
-
2008
- 2008-08-07 CN CN2008101182088A patent/CN101333961B/en active Active
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102337979A (en) * | 2011-08-11 | 2012-02-01 | 浙江大学 | Automatic calibration parameter optimization method of engine based on genetic algorithm |
| CN102626557A (en) * | 2012-04-13 | 2012-08-08 | 长春工业大学 | Molecular distillation process parameter optimizing method based on GA-BP (Genetic Algorithm-Back Propagation) algorithm |
| CN102717800A (en) * | 2012-06-13 | 2012-10-10 | 吉林大学 | Vehicle economy cruise control method based on engine universal characteristic |
| CN102717800B (en) * | 2012-06-13 | 2015-06-10 | 吉林大学 | Vehicle economy cruise control method based on engine universal characteristic |
| CN104598654A (en) * | 2014-10-07 | 2015-05-06 | 芜湖扬宇机电技术开发有限公司 | Ignition advance angle prediction system and method thereof |
| CN105020032A (en) * | 2015-07-31 | 2015-11-04 | 毛志明 | Novel self-adaptation gas fuel control system |
| CN106224088A (en) * | 2016-08-23 | 2016-12-14 | 石家庄新华能源环保科技股份有限公司 | A kind of method utilizing High Pressure Hydrogen fuel combination to drive automobile and dynamical system |
| CN106224088B (en) * | 2016-08-23 | 2019-03-08 | 石家庄新华能源环保科技股份有限公司 | A kind of method and dynamical system using High Pressure Hydrogen fuel combination driving automobile |
| CN108757264B (en) * | 2018-04-26 | 2019-12-10 | 杭州电子科技大学 | Method for obtaining optimal ignition advance angle of coke oven gas engine |
| CN108757264A (en) * | 2018-04-26 | 2018-11-06 | 杭州电子科技大学 | A method of obtaining coke-oven gas engine Optimum spark advance angle |
| CN108763681A (en) * | 2018-05-16 | 2018-11-06 | 华北水利水电大学 | Hydrogen engine fault diagnosis system and method based on FOA-GRNN blending algorithms |
| CN110080896A (en) * | 2019-04-24 | 2019-08-02 | 河南省图天新能源科技有限公司 | A kind of methane fuelled engine air/fuel ratio control method based on genetic algorithm |
| US11174800B2 (en) | 2019-09-24 | 2021-11-16 | Caterpillar Inc. | Transient controller and method of operating gas engine |
| CN112257355A (en) * | 2020-10-30 | 2021-01-22 | 东南大学 | Modeling method for medium-low pressure gas distribution pipe network of hydrogen-doped natural gas |
| CN113337321A (en) * | 2021-05-17 | 2021-09-03 | 西安交通大学 | Use of natural gas as hydrogen flame retardant and fuel suitable for oxyhydrogen burner |
| CN114483341A (en) * | 2022-01-12 | 2022-05-13 | 清华大学 | Combustion control method in hydrogen-doped natural gas engine |
| CN114961942A (en) * | 2022-06-14 | 2022-08-30 | 潍柴动力股份有限公司 | Method and device for reducing methane emission |
| CN114961942B (en) * | 2022-06-14 | 2024-04-16 | 潍柴动力股份有限公司 | Method and device for reducing methane emission |
| CN115324698A (en) * | 2022-08-09 | 2022-11-11 | 上海交通大学 | Diesel engine tail gas analysis sensing method based on neural network |
| CN115324698B (en) * | 2022-08-09 | 2023-06-16 | 上海交通大学 | Diesel engine tail gas analysis sensing method based on neural network |
| CN116006374A (en) * | 2023-02-20 | 2023-04-25 | 潍柴动力股份有限公司 | Fuel hydrogen-adding engine control method and device and vehicle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101333961B (en) | 2011-12-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101333961B (en) | Hydrogen gas natural gas mixed fuel engine optimizing method | |
| Brodny et al. | Assessing sustainable energy development in the central and eastern European countries and analyzing its diversity | |
| Menzel et al. | Multi-objective optimization of the volumetric and thermal efficiencies applied to a multi-cylinder internal combustion engine | |
| CN105116730B (en) | Hydrogen-fuel engine electronic spark advance angle and optimizing system and its optimization method based on Particle Group Fuzzy Neural Network | |
| Koç et al. | Prediction of emission and performance characteristics of reactivity-controlled compression ignition engine with the intelligent software based on adaptive neural-fuzzy and neural-network | |
| Kakaee et al. | Combining artificial neural network and multi-objective optimization to reduce a heavy-duty diesel engine emissions and fuel consumption | |
| Liu et al. | Optimization based on genetic algorithms on energy conservation potential of a high speed SI engine fueled with butanol–gasoline blends | |
| Dang et al. | Optimization control of hydrogen engine ignition system based on ACO-BP | |
| Roy et al. | ANN metamodel assisted Particle Swarm Optimization of the performance-emission trade-off characteristics of a single cylinder CRDI engine under CNG dual-fuel operation | |
| CN103870878B (en) | A kind of Combustion Characteristics in Utility Boiler neural network model | |
| Ma et al. | Multi-objective optimization of dual-fuel engine performance in PPCI mode based on preference decision | |
| CN107476903A (en) | A kind of supercharged diesel engine EGR Performance Evaluations and Optimum EGR rate determine method | |
| Mahla et al. | Optimization of engine operating variables on performance and emissions characteristics of biogas fuelled CI engine by the design of experiments: Taguchi approach | |
| Kumar et al. | Application of statistical approaches in IC engine calibration to enhance the performance and emission Characteristics: A methodological review | |
| Gokul Karthik et al. | Bald eagle search optimization on dual fueled reactivity controlled combustion ignition based engine characteristics by altering low reactive fuels | |
| CN113935124B (en) | Multi-target performance optimization method for biodiesel for diesel engine | |
| Manivasagam et al. | Investigation on 1-Propanol Electronic mode of fumigation on diesel engine performance and emission Fueled with diesel and lemongrass biodiesel blend using AHP-COPRAS | |
| Ming et al. | Supercritical thermophysical properties prediction of multi-component hydrocarbon fuels based on artificial neural network models | |
| Zhang et al. | Modelling and multi-objective combustion optimization of marine engine with speed maintaining control target | |
| CN108757228A (en) | A kind of supercharged diesel engine EGR performance estimating methods | |
| Dahake et al. | Experimental exploration of hydrogen fuelled dual fuelled CRDI equipped diesel engine coupled with exhaust gas recirculation using ANN approach | |
| CN108563844A (en) | A kind of supercharged diesel engine Optimum EGR rate determines method | |
| Shojaeefard et al. | Artificial neural network based multi-objective evolutionary optimization of a heavy-duty diesel engine | |
| CN113705890B (en) | Diesel engine emission pollutant control method based on approximate model | |
| CN114934848A (en) | Fuzzy neural network modeling method for optimizing control of combustion performance of diesel engine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |