CN118260975A - Engine combustion chamber numerical simulation method based on self-adaptive pre-zoning mechanism - Google Patents
Engine combustion chamber numerical simulation method based on self-adaptive pre-zoning mechanism Download PDFInfo
- Publication number
- CN118260975A CN118260975A CN202410693011.6A CN202410693011A CN118260975A CN 118260975 A CN118260975 A CN 118260975A CN 202410693011 A CN202410693011 A CN 202410693011A CN 118260975 A CN118260975 A CN 118260975A
- Authority
- CN
- China
- Prior art keywords
- state space
- state
- combustion chamber
- parameter correction
- space
- 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
- 230000007246 mechanism Effects 0.000 title claims abstract description 125
- 238000000034 method Methods 0.000 title claims abstract description 97
- 238000004088 simulation Methods 0.000 title claims abstract description 52
- 238000002485 combustion reaction Methods 0.000 title claims abstract description 44
- 238000013316 zoning Methods 0.000 title claims abstract description 19
- 238000006243 chemical reaction Methods 0.000 claims abstract description 99
- 238000004364 calculation method Methods 0.000 claims abstract description 46
- 238000005192 partition Methods 0.000 claims abstract description 6
- 238000010206 sensitivity analysis Methods 0.000 claims description 29
- 238000005457 optimization Methods 0.000 claims description 19
- 230000003044 adaptive effect Effects 0.000 claims description 15
- 230000035945 sensitivity Effects 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000006757 chemical reactions by type Methods 0.000 claims description 5
- 230000009191 jumping Effects 0.000 claims description 4
- 238000003064 k means clustering Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract description 4
- 230000011218 segmentation Effects 0.000 abstract description 3
- 230000000875 corresponding effect Effects 0.000 description 35
- 239000000446 fuel Substances 0.000 description 15
- 230000001133 acceleration Effects 0.000 description 13
- 230000000694 effects Effects 0.000 description 10
- 239000007789 gas Substances 0.000 description 9
- 239000000523 sample Substances 0.000 description 9
- UFHFLCQGNIYNRP-UHFFFAOYSA-N Hydrogen Chemical compound [H][H] UFHFLCQGNIYNRP-UHFFFAOYSA-N 0.000 description 7
- 230000002596 correlated effect Effects 0.000 description 7
- 239000001257 hydrogen Substances 0.000 description 7
- 229910052739 hydrogen Inorganic materials 0.000 description 7
- 239000004215 Carbon black (E152) Substances 0.000 description 5
- 229930195733 hydrocarbon Natural products 0.000 description 5
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 5
- 239000003054 catalyst Substances 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000002441 reversible effect Effects 0.000 description 4
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 3
- 150000002430 hydrocarbons Chemical class 0.000 description 3
- 239000001301 oxygen Substances 0.000 description 3
- 229910052760 oxygen Inorganic materials 0.000 description 3
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 2
- IMNFDUFMRHMDMM-UHFFFAOYSA-N N-Heptane Chemical compound CCCCCCC IMNFDUFMRHMDMM-UHFFFAOYSA-N 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 239000012530 fluid Substances 0.000 description 2
- -1 small molecule hydrocarbon Chemical class 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 229910052757 nitrogen Inorganic materials 0.000 description 1
- 239000007800 oxidant agent Substances 0.000 description 1
- 230000001590 oxidative effect Effects 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 239000000376 reactant Substances 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
- 239000013074 reference sample Substances 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/12—Timing analysis or timing optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an engine combustion chamber numerical simulation method based on a self-adaptive pre-zoning mechanism, which comprises the following steps of: s1, determining a reaction formula with obvious influence on ignition delay time; s2, constructing a reactive state space; s3, storing the optimized parameter correction coefficients into the corresponding state space in each current highest-level state space; s4, determining a state space needing encryption; s5, encrypting all the state spaces needing to be encrypted; s6, storing the parameter correction coefficients in the encrypted state space; and S7, searching the parameter correction coefficient in the combustion chamber numerical simulation calculation process. The invention solves the defects existing in the self-adaptive pre-segmentation mechanism before, realizes the simplicity and the high efficiency of state space partition retrieval, thereby effectively improving the calculation efficiency and the accuracy of the current numerical simulation program on the numerical simulation of the engine combustion chamber, and has strong applicability and application space.
Description
Technical Field
The invention belongs to the technical field of computer simulation of combustion processes of aero-engine combustion chambers, and particularly relates to an engine combustion chamber numerical simulation method based on a self-adaptive pre-zoning mechanism.
Background
As modern aircraft have increased demands on engine performance, the problem of numerical modeling of the combustion flow process within the engine combustion chamber has become a research hotspot. When the process is numerically simulated, the reaction mechanism of the fuel plays a vital role in the calculation cost, calculation efficiency and calculation accuracy of the numerical simulation. In general, reaction mechanisms include a detailed mechanism, a backbone mechanism, a simplified mechanism, and a general packet mechanism, each of which has merits and merits with each other. The detailed mechanism includes, in addition to the fuel and oxidant (typically oxygen-based), a large number of intermediate components and primitive reactions, which themselves are computationally intensive and greatly increase the computational effort of numerical modeling (generally considered to be positively correlated with the number of components and negatively correlated with the time advance step available). The backbone mechanism is based on the detailed mechanism, and the number of components is reduced by deleting unimportant elementary reactions and intermediate components. The simplified mechanism is further provided, and the application range of the mechanism is narrowed, so that elementary reactions and intermediate components which are not important in a specific application range are further deleted and combined, and the number of the components is further reduced. As for the total package mechanism, only a few intermediate components are contained, and the total package reaction is used to describe the combustion process of the fuel, resulting in a more general accuracy. The characteristics of the four mechanisms can be summarized by using fig. 1, and the ideal mechanism applicable to numerical simulation has both calculation efficiency and accuracy.
The reaction mechanism actually used in computational fluid dynamics (Computational Fluid Dynamics, CFD) is basically a simplified mechanism, but the simplified mechanism cannot be kept consistent with the detailed mechanism calculation result under a wide range of reaction conditions, especially under low temperature and low pressure conditions. To get as close to the effect of the ideal mechanism as possible Schwer first proposed the concept of an adaptive chemical approach in 2003, i.e. invoking a corresponding simplified mechanism based on the operating conditions to accelerate the computation. Schwer generates 4 simplified mechanisms based on the detailed reaction mechanism of hydrogen: oxygen enrichment, hydrogen enrichment, H2/N2 mixing and H2/O2 mixing, and switching corresponding reaction mechanisms according to the mass fraction of the components, and applying the reaction mechanisms to calculating a hydrogen shear layer calculation example, the acceleration of 2-3 times is found to be realized. The state space division method used by the method is more original, and the simplification of the mechanism is not discussed deeply, but a concept (Pre-Partition ADAPTIVE CHEMISTRY, PPAC) of a Pre-region self-adaptive reaction mechanism is created, so that the calculation amount of the application of the simplification mechanism can accurately calculate the chemical reaction in a wide range. Oluwol and Banerjee subsequently studied the effective domain and search problems of the simplified mechanism and were successfully applied to the simple one-dimensional methane premixed flame example. Liang further developed a search method for mechanism simplification and a preconfigured mechanism library, namely, a nearest neighbor search algorithm is adopted to search the reaction mechanism, and finally, the reaction mechanism is searched according to GRI 3.0 (GRI 3.0 is a chemical kinetics mechanism for describing the combustion process of hydrocarbon fuel). This mechanism was developed by the Institute GAS RESEARCH Institute of natural gas, which included a large number of chemical reaction steps and reactants to more accurately simulate the combustion behavior of methane under different conditions) generated 18 simplified mechanisms, and tested the simplified mechanisms in a homogeneous charge compression ignition engine simulation, saving 50% of the computation time. However, the pre-zoning adaptive mechanism at the moment has application problems including complexity of algorithm implementation and application and completeness of state space estimation, and Liang proposes a dynamic adaptive chemistry concept (DYNAMIC ADAPTIVE CHEMISTRY, DAC), namely, real-time simplification of a detailed reaction mechanism in a CFD simulation process. The core idea is to generate a simplified mechanism in each grid according to local thermodynamic state parameters and directly use the mechanism, so that estimation of state space and retrieval are not needed. Liang also accelerated 30-fold using this method in a homogeneous charge compression ignition engine simulation with heptane (detailed mechanism 578 component) fuel. Although DAC-like methods have been published in the last decade with a series of important research results, these methods have a common problem: the acceleration effect is limited on detailed reaction mechanisms with fewer components. For example, at GRI 3.0, the 53 component speed ratio is only between 3-4 times. With the current computing power, the real aero-engine combustion chamber numerical simulation still cannot use a detailed reaction mechanism (unless a flame surface model is adopted), and even if a DAC method is adopted, a good acceleration effect is difficult to obtain. in addition, the real-time simplification method brings great difficulty to program parallelism because of great difference of calculation efficiency under different states.
Disclosure of Invention
Aiming at the defects in the prior art, the engine combustion chamber numerical simulation method based on the self-adaptive pre-segmentation mechanism solves the problem that the existing self-adaptive pre-segmentation mechanism cannot achieve both calculation efficiency and accuracy.
In order to achieve the aim of the invention, the invention adopts the following technical scheme: an engine combustion chamber numerical simulation method based on a self-adaptive pre-zoning mechanism comprises the following steps:
s1, performing sensitivity analysis on each reaction formula in a simplified mechanism, and determining a reaction formula with obvious influence on ignition delay time;
S2, based on the state of the self-ignition simulation of the reaction formula with obvious influence on the ignition delay time, constructing a state space of a reaction mechanism, and setting the dividing number of the state space range and the 0 th-level state space;
S3, in each current highest-level state space, performing mechanism parameter optimization by using a parameter correction coefficient of a previous level as an optimization initial value, and storing the optimized parameter correction coefficient into a corresponding state space;
s4, obtaining the local maximum error of the optimized parameter correction coefficient in each state space, and determining the state space needing encryption;
S5, judging whether a state space needing encryption exists, if so, encrypting all the state spaces needing encryption to generate a state space with a higher level, returning to S3, and if not, entering S6;
s6, storing parameter correction coefficients in all the hierarchical state spaces after the encryption process is completed in a list form, and storing the hierarchical affiliations in a database in a jump sequence number mode;
And S7, searching parameter correction coefficients stored in a list form according to the local state of the calculated grid in the combustion chamber numerical simulation calculation process, and applying the searched parameter correction coefficients to the combustion chamber numerical simulation calculation.
Further: the step S1 comprises the following sub-steps:
s11, determining the position of a sample point required by sensitivity analysis according to a reaction formula in a simplified mechanism;
S12, setting a state of performing self-ignition simulation in a reaction mode;
S13, performing self-ignition simulation on the premixed gas through mechanism coefficients corresponding to the sensitivity analysis sampling points to obtain ignition delay time obtained through simulation under the mechanism coefficients corresponding to the sensitivity analysis sampling points in different states;
S14, performing sensitivity analysis on the sensitivity analysis sample points in each state and the corresponding ignition delay time, so as to obtain sensitivity analysis results in the corresponding states;
S15, taking sensitivity analysis results of all reaction formulas in different states as 5-dimensional sensitivity vectors corresponding to all the reaction formulas, and carrying out k-means clustering on the 5-dimensional sensitivity vectors to obtain a first cluster and a second cluster, wherein the average sensitivity of the first cluster is higher than that of the second cluster;
s16, taking the reaction formula in the first cluster as the reaction formula with obvious influence on the ignition delay time.
Further: in S12, the method for setting the state of the reactive type to perform the auto-ignition simulation specifically includes:
Determining high, medium and low pressure values in a pressure range according to the upper and lower limits of the input pressure;
And determining high, medium and low temperature values in a temperature range according to the upper and lower limits of the input temperature, and combining the pressure range and the temperature range to obtain a state of performing self-ignition simulation of the reaction formula.
Further: in the step S3, the method for optimizing the mechanism parameters specifically comprises the following steps:
Optimizing by using a Nelder-Mead simplex optimization algorithm and taking the initial value of the parameter correction coefficient in the corresponding state space as an optimization initial value and the relative error of the ignition delay time of the modified mechanism as a minimum optimization target to obtain an optimized parameter correction coefficient;
The calculation mode of the ignition delay time relative error specifically comprises the following steps: and taking the state corresponding to the state space center point as an initial state, performing zero-dimensional self-ignition simulation by using a Cantera function library by using a detailed mechanism and a simplified mechanism after changing parameters, and calculating the relative error of the ignition delay time corresponding to the modified parameters.
Further: in the step S4, the method for determining the state space to be encrypted includes:
If the local maximum error of the optimized parameter correction coefficient in the state space is larger than a set threshold value, the state space is used as the state space needing encryption;
the method for calculating the local maximum error of the optimized parameter correction coefficient in any state space specifically comprises the following steps:
And 8 state points uniformly distributed in the state space are selected as initial points, the relative error of the optimized parameter correction coefficient on the state points is calculated according to the calculation mode of the relative error of the ignition delay time, and the maximum relative error is taken as the local maximum error.
Further: in the step S5, the method for encrypting the state space specifically includes:
and halving the state space on each dimension to generate 8 sub-state spaces, and marking the hierarchy of the sub-state spaces as the hierarchy of the original state space plus 1, wherein the sub-state spaces are the same as the original state space, and the space size is 1/8 of the original state space.
Further: in the step S7, the searching method specifically includes:
S71, normalizing the state in the current calculation unit, and calculating the relative position of the current calculation unit in a state space represented by a database;
S72, jumping to the corresponding running data reading of the 0 th-level state space according to the relative position;
s73, judging whether a higher-level sub-state space exists in the current state space, if so, entering S74; if not, entering S75;
S74, calculating offset according to the position of the read data in the current state space, entering a corresponding next-level state space according to the offset, and returning to S73;
And S75, stopping the search and outputting the parameter correction coefficients read in the current state space.
Further: in S74, the method for calculating the offset is as follows:
Starting from the 0 th-level state space, setting the reference values of the state space and the sub-state space according to the sequence numbers of the 0 th-level state space and the sub-state space, taking the relative position as the lower left sub-state space as a reference value +1, the sequence number at the lower right is the reference value +2, the sequence number at the upper left is the reference value +3, obtaining the read data reference value according to the position of the read data in the current state space, and adding the read data reference value and the reference value of the current state space to obtain the offset.
The beneficial effects of the invention are as follows:
(1) The invention provides a numerical simulation method of an engine combustion chamber based on a self-adaptive pre-zoning mechanism, which solves the defects existing in the self-adaptive pre-zoning mechanism before, realizes the simplicity and the high efficiency of state space zoning search, thereby effectively improving the calculation efficiency and the accuracy of the current numerical simulation program on the numerical simulation of the engine combustion chamber; in addition, although the method describes a finite state space, explicit search logic is provided for the state points beyond the state space, so that any state point can find the subspace and the corresponding pre-finger factor adjustment coefficient in the database of the method.
(2) Compared with a DAC method, the method does not need to pay extra calculation cost in the CFD calculation process to simplify the mechanism or optimize the mechanism, and only needs to pay the layer-by-layer jump time of the multi-layer octree, so that the calculation cost is greatly reduced. The DAC method needs to perform mechanism simplification in real time in the CFD calculation process, and although the used mechanism simplification methods such as DRG (Directed relation graph, direct relation diagram) and the like do have very high calculation efficiency, it is difficult to perform mechanism parameter optimization in comparison with the state space of the construction reaction type of the present invention.
(3) The method has strong applicability and application space, can obtain better acceleration effect on detailed reaction mechanism with less components, and compared with a DAC method, the acceleration effect presented by the DAC method is more than represented on complex hydrocarbon fuel with the component number of 100+, so that the acceleration effect of tens times or even tens times is realized, but the application report on hydrogen fuel or small molecule hydrocarbon fuel with smaller reaction scale is rare. In principle, the limit of acceleration capability is the level of tens of reactions of tens of components. While the acceleration capability limit of the present method depends on the proper simplification mechanism, the minimum simplification mechanism of common fuels in general can reach levels below ten components. Considering that the calculated amount of the non-combustion part in CFD calculation is positively correlated with the component number, the calculated amount of the combustion part is positively correlated with the component number and the reaction number, the upper limit of the accelerating capacity and the application range of the method of the invention obviously exceed the DAC method.
Drawings
FIG. 1 is a schematic diagram showing the comparison of the characteristics of the four mechanisms.
FIG. 2 is a flow chart of a method for modeling engine combustion chamber values based on an adaptive pre-zoning mechanism of the present invention.
Fig. 3 is a two-dimensional cut-away view of a three-dimensional adaptively encrypted state space.
Fig. 4 is a schematic diagram of the retrieval of the state space of adaptive encryption.
Fig. 5 is a schematic diagram of a storage form of the parameter correction coefficients in the database.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 2, in one embodiment of the invention, an engine combustion chamber numerical simulation method based on an adaptive pre-zoning mechanism includes the steps of:
s1, performing sensitivity analysis on each reaction formula in a simplified mechanism, and determining a reaction formula with obvious influence on ignition delay time;
S2, based on the state of the self-ignition simulation of the reaction formula with obvious influence on the ignition delay time, constructing a state space of a reaction mechanism, and setting the dividing number of the state space range and the 0 th-level state space;
S3, in each current highest-level state space, performing mechanism parameter optimization by using a parameter correction coefficient of a previous level as an optimization initial value, and storing the optimized parameter correction coefficient into a corresponding state space;
s4, obtaining the local maximum error of the optimized parameter correction coefficient in each state space, and determining the state space needing encryption;
S5, judging whether a state space needing encryption exists, if so, encrypting all the state spaces needing encryption to generate a state space with a higher level, returning to S3, and if not, entering S6;
s6, storing parameter correction coefficients in all the hierarchical state spaces after the encryption process is completed in a list form, and storing the hierarchical affiliations in a database in a jump sequence number mode;
And S7, searching parameter correction coefficients stored in a list form according to the local state of the calculated grid in the combustion chamber numerical simulation calculation process, and applying the searched parameter correction coefficients to the combustion chamber numerical simulation calculation.
The basic idea of the invention is to carry out self-adaptive adjustment on parameters in a simplified mechanism of a fixed form under different states, realize self-adaptive mechanisms of the fixed form and the self-adaptive change of the parameters, and the current self-adaptive mechanisms are basically the form change and the parameter fixation, and ensure that the combustion characteristics and detailed mechanisms of the self-adaptive mechanisms are consistent under a wide range.
The step S1 comprises the following sub-steps:
s11, determining the position of a sample point required by sensitivity analysis according to a reaction formula in a simplified mechanism;
in the present embodiment, the number of sample points and the position are correlated with the dimension for which the sensitivity analysis is directed, and the number of sample points is determined from the set reference number of sample points N and the dimension D, wherein the former needs to be a power of 2. In the invention, the number of reference sample points is 1024, taking a hydrogen ES mechanism as an example, 16 reactions exist, the total sample number is N (2D+2) =1024 (2×16+2) =34816, the sample point positions are Saltelli samples based on a Sobol sequence, and the method belongs to a mature algorithm in the sensitivity analysis field.
S12, setting a state of performing self-ignition simulation in a reaction mode;
S13, performing self-ignition simulation on the premixed gas through mechanism coefficients corresponding to the sensitivity analysis sampling points to obtain ignition delay time obtained through simulation under the mechanism coefficients corresponding to the sensitivity analysis sampling points in different states;
In this embodiment, the sensitivity analysis adopts a Sobol sensitivity analysis method for performing sensitivity analysis on the pre-finger factors of each reaction formula in the simplified mechanism, so as to determine which of the pre-finger factors of the reaction formula are adjusted to most effectively influence the combustion characteristic of the ignition delay time of the mechanism.
The premix gas composition is combined with the corresponding oxygen and nitrogen in equivalent mixing according to the entered fuel name and corresponding stoichiometric number.
S14, performing sensitivity analysis on the sensitivity analysis sample points in each state and the corresponding ignition delay time, so as to obtain sensitivity analysis results in the corresponding states;
S15, taking sensitivity analysis results of all reaction formulas in different states as 5-dimensional sensitivity vectors corresponding to all the reaction formulas, and carrying out k-means clustering on the 5-dimensional sensitivity vectors to obtain a first cluster and a second cluster, wherein the average sensitivity of the first cluster is higher than that of the second cluster;
s16, taking the reaction formula in the first cluster as the reaction formula with obvious influence on the ignition delay time.
In S12, the method for setting the state of the reactive type to perform the auto-ignition simulation specifically includes:
Determining high, medium and low pressure values in a pressure range according to the upper and lower limits of the input pressure;
And determining high, medium and low temperature values in a temperature range according to the upper and lower limits of the input temperature, and combining the pressure range and the temperature range to obtain a state of performing self-ignition simulation of the reaction formula.
It includes the states of high temperature and high pressure, high temperature and low pressure, low temperature and high pressure, low temperature and low pressure and medium temperature and medium pressure.
In the case of the reaction formula in which hydrogen is used as a fuel in S2, the temperature ranges from 1000K to 1800K, the pressure ranges from 0.5 to atm to 10 atm, and the equivalence ratio ranges from 0.5 to 2.0. The 0 th-level state space is divided into four equal parts in each dimension, so that 64 0 th-level state spaces are obtained in total, and the initial value of the parameter correction coefficient of each reaction type pre-finger factor is set to be 1.0.
In this embodiment, the principle of the present invention for constructing the state space of the reaction formula is specifically:
The chemical reaction process can be described by the following set of ordinary differential equations:
Wherein the method comprises the steps of Is the firstThe mass of the individual components is determined,It is the time that is required for the device to be in contact with the substrate,Is the first corresponding to the chemical reactionThe formation rate of each component:
Wherein the method comprises the steps of The number of the reactions is the number of the reactions,Is the number of the components of the composition,Is the firstThe molar mass of the individual components is determined,Is the firstIn the first reactionThe amount of change in the stoichiometry of the individual components,Is the firstThe positive reaction rate constant of each reaction,Is the firstThe inverse reaction rate constant of each reaction,Is the firstThe molar concentration of the individual components is determined,Is the firstIn the first reactionPositive reaction concentration index of the individual components,Is the firstIn the first reactionReverse reaction concentration index of each component.
For a general primitive reaction, the forward and reverse reaction concentration index is consistent with the stoichiometric number of the reaction formula, for 2H 2+O2<=>H2O,H2, the forward reaction concentration index of O 2 is 2, the reverse reaction concentration index of H 2 O is 2, and the other concentration indexes are 0. However, in the case of a relatively special case, the reaction concentration index may be additionally specified.
Generally, each chemical reaction formula has three parameters to describe the positive reaction rate for that reaction formulaIs calculated by the pre-finger factor in Alnius equation (Arrhenius equation)Index of temperatureAnd activation energy：
In the method, in the process of the invention,Is the positive reaction rate and the negative reaction rate,It is meant that the pro-factor is,Is the temperature measured in thermodynamic units,Is the concentration index, exp is an exponential function,Is the activation energy of the catalyst and is used for preparing the catalyst,As an ideal gas constant, 8.314J/(mol.K) is generally used.
The general elementary reactions are all reversible reactions, for which the rate of the reaction is reversedGenerally by reaction equilibrium constantAnd (3) calculating to obtain:
In the method, in the process of the invention, Is the reaction rate of the reverse reaction, and the reaction rate of the catalyst is the same as that of the catalyst,Is the equilibrium constant of the reaction, and the reaction time is,Is at standard atmospheric pressure, typically 101325 Pa,The number of the components is a plurality of,Is the first side of the raw materialThe stoichiometry of the individual components,Is the product sideThe stoichiometry of the individual components, exp, represents an exponential function,Represent the firstThe entropy of the individual components in the standard state,Represent the firstThe gas constants of the individual components are set,，Is the firstThe molar mass of the individual components is determined,Represent the firstEnthalpy value of the individual components at the current temperature.
By varying these parameters, the combustion performance exhibited by the mechanism will also change. Therefore, the invention adjusts parameters through an optimization algorithm, so that the adjusted simplified mechanism has combustion characteristics consistent with the detailed mechanism in any state. Any state herein refers to any single state, not any state, which means that in different states, corresponding, suitable parameters can be optimized by an optimization algorithm, but not identical to each other. In the present invention, three physical quantities of gas temperature, gas pressure, and gas equivalence ratio are used as dimensions of the state space in consideration of the use effect and the computational convenience.
In order to solve the problems of complex state space and difficult retrieval of the pre-partition self-adaptive mechanism, the method stores and calls the state space of the three-dimensional rectangular self-adaptive encryption form shown in fig. 3, wherein the abscissa is pressure and the ordinate is temperature. In the state space of the form, a plurality of layers of square state spaces exist, the shapes of the square state spaces in the layers are the same, each low-level state space can be uniformly divided into 8 (2 x 2) higher-level state spaces, and the state space of the lowest level is a cuboid shape with a three-dimensional space structure and uniform division. Therefore, in the state space, any state point can carry out layer-by-layer judgment on each layer position to which the state point belongs, so that the minimum state space to which the state point belongs is found. The searching process can refer to the process shown in fig. 4, wherein the target state point X and the self-adaptive encryption state space 0-8, the state space 0 is the 0 th level, the state space 1-4 is the 1 st level, the state space 5-8 is the 2 nd level, and the state space to which the state point X belongs is searched step by step according to the sequence of 0-3-6. For a state point that is outside the state space range, it means that at least one of the three physical quantities is outside the delimited range, is adjusted to be the nearest in-range point, and for a larger-than-upper-bound to be an upper-bound and a smaller-than-lower-bound to be a lower-bound, the components that are in-range remain unchanged.
In the step S3, the method for optimizing the mechanism parameters specifically comprises the following steps:
Optimizing by using a Nelder-Mead simplex optimization algorithm and taking the initial value of the parameter correction coefficient in the corresponding state space as an optimization initial value and the relative error of the ignition delay time of the modified mechanism as a minimum optimization target to obtain an optimized parameter correction coefficient;
The calculation mode of the ignition delay time relative error specifically comprises the following steps: and taking the state corresponding to the state space center point as an initial state, performing zero-dimensional self-ignition simulation by using a Cantera function library by using a detailed mechanism and a simplified mechanism after changing parameters, and calculating the relative error of the ignition delay time corresponding to the modified parameters.
In this embodiment, the mechanism parameter to be optimized is specifically a pre-finger factor of a reaction type that affects the ignition delay time significantly.
In the step S4, the method for determining the state space to be encrypted includes:
If the local maximum error of the optimized parameter correction coefficient in the state space is larger than a set threshold value, the state space is used as the state space needing encryption; in this embodiment, the threshold value is set to 10%.
The method for calculating the local maximum error of the optimized parameter correction coefficient in any state space specifically comprises the following steps:
And 8 state points uniformly distributed in the state space are selected as initial points, the 8 state points are obtained by using a Sobol sequence, the relative errors of the optimized parameter correction coefficients on the state points are calculated according to the calculation mode of the relative errors of the ignition delay time, and the maximum relative error is taken as a local maximum error.
In the step S5, the method for encrypting the state space specifically includes:
and halving the state space on each dimension to generate 8 sub-state spaces, and marking the hierarchy of the sub-state spaces as the hierarchy of the original state space plus 1, wherein the sub-state spaces are the same as the original state space, and the space size is 1/8 of the original state space.
In S5, if there is no state space that needs to be encrypted, this means that the maximum relative error estimated in all the minimum state spaces is smaller than the set threshold.
In the step S6, the storage form of the parameter correction coefficients in the database is shown in fig. 5, where the first column is the current level, the second column is the jump target sequence number, and the third to fifth columns are the pre-finger factor correction coefficients of three main reactions.
In the step S7, the searching method specifically includes:
S71, normalizing the state in the current calculation unit, and calculating the relative position of the current calculation unit in a state space represented by a database;
S72, jumping to the corresponding running data reading of the 0 th-level state space according to the relative position;
s73, judging whether a higher-level sub-state space exists in the current state space, if so, entering S74; if not, entering S75;
S74, calculating offset according to the position of the read data in the current state space, entering a corresponding next-level state space according to the offset, and returning to S73;
in S74, the method for calculating the offset is as follows:
Starting from the 0 th-level state space, setting the reference values of the state space and the sub-state space according to the sequence numbers of the 0 th-level state space and the sub-state space, taking the relative position as the lower left sub-state space as a reference value +1, the sequence number at the lower right is the reference value +2, the sequence number at the upper left is the reference value +3, obtaining the read data reference value according to the position of the read data in the current state space, and adding the read data reference value and the reference value of the current state space to obtain the offset.
In this embodiment, the method for calculating the offset is specifically: as shown in fig. 4, starting from the 0 th-level state space, the reference value of the state space and the sub-state space thereof is set according to the relative position of the state point in the upper-level state space, the sub-state space with the relative position at the lower left is used as the reference value +1, the reference value at the upper left is used as the reference value +2, the reference value at the upper right is used as the reference value +3, and the offset of the sub-state space sequence number to which the state point belongs relative to the reference value can be known according to the relative position of the state point in the upper-level state space.
As shown in fig. 4, the target state point of the yoke is located at the upper left of the stage state space 0, where the reference value is 2, and the reference value of the sub-state space of the state space o is 1, so the resulting offset is: 1+2=3, the number of the sub-state space where the target state point is located is 3, so as to jump to the state space 3, then, in the state space 3, the target state point is located at the lower right thereof, the corresponding reference value is 1, and the reference value of the sub-state space of the state space 3 is 5, so as to obtain that the target state point finally falls in the state space 6.
The above description is generalized in the three-dimensional state space used in the present invention, and whether the binary offset corresponding bit is 0 or 1 is set with the relative magnitude of the state space center value. The first dimension has a smaller central value, and the first bit of the binary offset is 0; the larger, the first bit of the binary offset is 1. By analogy, the relative positions of all the sub-state spaces in three dimensions and their corresponding offsets from the reference value can be obtained. And according to the state in the current calculation unit, the relative position corresponding to the current calculation unit can be judged to be compared with the central value of the state space, and the offset is obtained according to the corresponding concern described above, so that the sequence number of the next layer space in which the current calculation unit is positioned, namely the number of the jumping lines and the offset are obtained.
And S75, stopping the search and outputting the parameter correction coefficients read in the current state space.
The beneficial effects of the invention are as follows: the invention provides a numerical simulation method of an engine combustion chamber based on a self-adaptive pre-zoning mechanism, which solves the defects existing in the self-adaptive pre-zoning mechanism before, realizes the simplicity and the high efficiency of state space zoning search, thereby effectively improving the calculation efficiency and the accuracy of the current numerical simulation program on the numerical simulation of the engine combustion chamber; in addition, although the method describes a finite state space, explicit search logic is provided for the state points beyond the state space, so that any state point can find the subspace and the corresponding pre-finger factor adjustment coefficient in the database of the method.
Compared with a DAC method, the method does not need to pay extra calculation cost in the CFD calculation process to simplify the mechanism or optimize the mechanism, and only needs to pay the layer-by-layer jump time of the multi-layer octree, so that the calculation cost is greatly reduced. The DAC method needs to perform mechanism simplification in real time in the CFD calculation process, and the adopted mechanism simplification methods such as DRG and the like have high calculation efficiency, but are difficult to compare with the construction reaction type state space of the invention in mechanism parameter optimization.
The method has strong applicability and application space, can obtain better acceleration effect on detailed reaction mechanism with less components, and compared with a DAC method, the acceleration effect presented by the DAC method is more than represented on complex hydrocarbon fuel with the component number of 100+, so that the acceleration effect of tens times or even tens times is realized, but the application report on hydrogen fuel or small molecule hydrocarbon fuel with smaller reaction scale is rare. In principle, the limit of acceleration capability is the level of tens of reactions of tens of components. While the acceleration capability limit of the present method depends on the proper simplification mechanism, the minimum simplification mechanism of common fuels in general can reach levels below ten components. Considering that the calculated amount of the non-combustion part in CFD calculation is positively correlated with the component number, the calculated amount of the combustion part is positively correlated with the component number and the reaction number, the upper limit of the accelerating capacity and the application range of the method of the invention obviously exceed the DAC method.
In the description of the present invention, it should be understood that the terms "center," "thickness," "upper," "lower," "horizontal," "top," "bottom," "inner," "outer," "radial," and the like indicate or are based on the orientation or positional relationship shown in the drawings, merely to facilitate description of the present invention and to simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be configured and operated in a particular orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be interpreted as indicating or implying a relative importance or number of technical features indicated. Thus, a feature defined as "first," "second," "third," or the like, may explicitly or implicitly include one or more such feature.
Claims (8)
1. The engine combustion chamber numerical simulation method based on the self-adaptive pre-zoning mechanism is characterized by comprising the following steps of:
s1, performing sensitivity analysis on each reaction formula in a simplified mechanism, and determining a reaction formula with obvious influence on ignition delay time;
S2, based on the state of the self-ignition simulation of the reaction formula with obvious influence on the ignition delay time, constructing a state space of a reaction mechanism, and setting the dividing number of the state space range and the 0 th-level state space;
S3, in each current highest-level state space, performing mechanism parameter optimization by using a parameter correction coefficient of a previous level as an optimization initial value, and storing the optimized parameter correction coefficient into a corresponding state space;
s4, obtaining the local maximum error of the optimized parameter correction coefficient in each state space, and determining the state space needing encryption;
S5, judging whether a state space needing encryption exists, if so, encrypting all the state spaces needing encryption to generate a state space with a higher level, returning to S3, and if not, entering S6;
s6, storing parameter correction coefficients in all the hierarchical state spaces after the encryption process is completed in a list form, and storing the hierarchical affiliations in a database in a jump sequence number mode;
And S7, searching parameter correction coefficients stored in a list form according to the local state of the calculated grid in the combustion chamber numerical simulation calculation process, and applying the searched parameter correction coefficients to the combustion chamber numerical simulation calculation.
2. The engine combustion chamber numerical simulation method based on the adaptive pre-zoning mechanism according to claim 1, wherein S1 comprises the following sub-steps:
s11, determining the position of a sample point required by sensitivity analysis according to a reaction formula in a simplified mechanism;
S12, setting a state of performing self-ignition simulation in a reaction mode;
S13, performing self-ignition simulation on the premixed gas through mechanism coefficients corresponding to the sensitivity analysis sampling points to obtain ignition delay time obtained through simulation under the mechanism coefficients corresponding to the sensitivity analysis sampling points in different states;
S14, performing sensitivity analysis on the sensitivity analysis sample points in each state and the corresponding ignition delay time, so as to obtain sensitivity analysis results in the corresponding states;
S15, taking sensitivity analysis results of all reaction formulas in different states as 5-dimensional sensitivity vectors corresponding to all the reaction formulas, and carrying out k-means clustering on the 5-dimensional sensitivity vectors to obtain a first cluster and a second cluster, wherein the average sensitivity of the first cluster is higher than that of the second cluster;
s16, taking the reaction formula in the first cluster as the reaction formula with obvious influence on the ignition delay time.
3. The method for simulating the engine combustion chamber value based on the adaptive pre-zoning mechanism according to claim 2, wherein in S12, the method for setting the state of the reaction type for performing the auto-ignition simulation is specifically as follows:
Determining high, medium and low pressure values in a pressure range according to the upper and lower limits of the input pressure;
And determining high, medium and low temperature values in a temperature range according to the upper and lower limits of the input temperature, and combining the pressure range and the temperature range to obtain a state of performing self-ignition simulation of the reaction formula.
4. The engine combustion chamber numerical simulation method based on the adaptive pre-zoning mechanism according to claim 1, wherein in S3, the method for optimizing the mechanism parameters is specifically as follows:
Optimizing by using a Nelder-Mead simplex optimization algorithm and taking the initial value of the parameter correction coefficient in the corresponding state space as an optimization initial value and the relative error of the ignition delay time of the modified mechanism as a minimum optimization target to obtain an optimized parameter correction coefficient;
The calculation mode of the ignition delay time relative error specifically comprises the following steps: and taking the state corresponding to the state space center point as an initial state, performing zero-dimensional self-ignition simulation by using a Cantera function library by using a detailed mechanism and a simplified mechanism after changing parameters, and calculating the relative error of the ignition delay time corresponding to the modified parameters.
5. The engine combustion chamber numerical simulation method based on the adaptive pre-partition mechanism according to claim 4, wherein in S4, the method for determining the state space to be encrypted is as follows:
If the local maximum error of the optimized parameter correction coefficient in the state space is larger than a set threshold value, the state space is used as the state space needing encryption;
the method for calculating the local maximum error of the optimized parameter correction coefficient in any state space specifically comprises the following steps:
And 8 state points uniformly distributed in the state space are selected as initial points, the relative error of the optimized parameter correction coefficient on the state points is calculated according to the calculation mode of the relative error of the ignition delay time, and the maximum relative error is taken as the local maximum error.
6. The engine combustion chamber numerical simulation method based on the adaptive pre-partition mechanism according to claim 1, wherein in S5, the method for encrypting the state space specifically comprises:
and halving the state space on each dimension to generate 8 sub-state spaces, and marking the hierarchy of the sub-state spaces as the hierarchy of the original state space plus 1, wherein the sub-state spaces are the same as the original state space, and the space size is 1/8 of the original state space.
7. The engine combustion chamber numerical simulation method based on the adaptive pre-partition mechanism according to claim 1, wherein in S7, the searching method specifically comprises:
S71, normalizing the state in the current calculation unit, and calculating the relative position of the current calculation unit in a state space represented by a database;
S72, jumping to the corresponding running data reading of the 0 th-level state space according to the relative position;
s73, judging whether a higher-level sub-state space exists in the current state space, if so, entering S74; if not, entering S75;
S74, calculating offset according to the position of the read data in the current state space, entering a corresponding next-level state space according to the offset, and returning to S73;
And S75, stopping the search and outputting the parameter correction coefficients read in the current state space.
8. The method for modeling engine combustion chamber values based on the adaptive pre-zoning mechanism according to claim 7, wherein in S74, the method for calculating the offset is as follows:
Starting from the 0 th-level state space, setting the reference values of the state space and the sub-state space according to the sequence numbers of the 0 th-level state space and the sub-state space, taking the relative position as the lower left sub-state space as a reference value +1, the sequence number at the lower right is the reference value +2, the sequence number at the upper left is the reference value +3, obtaining the read data reference value according to the position of the read data in the current state space, and adding the read data reference value and the reference value of the current state space to obtain the offset.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410693011.6A CN118260975B (en) | 2024-05-31 | 2024-05-31 | Engine combustion chamber numerical simulation method based on self-adaptive pre-zoning mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410693011.6A CN118260975B (en) | 2024-05-31 | 2024-05-31 | Engine combustion chamber numerical simulation method based on self-adaptive pre-zoning mechanism |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118260975A true CN118260975A (en) | 2024-06-28 |
| CN118260975B CN118260975B (en) | 2024-08-06 |
Family
ID=91613440
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410693011.6A Active CN118260975B (en) | 2024-05-31 | 2024-05-31 | Engine combustion chamber numerical simulation method based on self-adaptive pre-zoning mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118260975B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200184131A1 (en) * | 2018-06-27 | 2020-06-11 | Dalian University Of Technology | A method for prediction of key performance parameter of an aero-engine transition state acceleration process based on space reconstruction |
| CN111762958A (en) * | 2020-05-12 | 2020-10-13 | 中铁第四勘察设计院集团有限公司 | Deep well aeration process optimization method and device for sewage treatment plant based on ASM2D model |
| CN114171127A (en) * | 2021-12-07 | 2022-03-11 | 上海交通大学四川研究院 | Method for simulating supersonic combustion flow by constructing self-adaptive reaction mechanism |
| US20220328139A1 (en) * | 2019-06-03 | 2022-10-13 | Sanofi Pasteur | Method for determining at a current time point a preservation state of one product and computer system for carrying out said method |
| WO2022256682A1 (en) * | 2021-06-03 | 2022-12-08 | The Texas A&M University System | Systems and methods for determining constitutive parameters of subject materials |
-
2024
- 2024-05-31 CN CN202410693011.6A patent/CN118260975B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200184131A1 (en) * | 2018-06-27 | 2020-06-11 | Dalian University Of Technology | A method for prediction of key performance parameter of an aero-engine transition state acceleration process based on space reconstruction |
| US20220328139A1 (en) * | 2019-06-03 | 2022-10-13 | Sanofi Pasteur | Method for determining at a current time point a preservation state of one product and computer system for carrying out said method |
| CN111762958A (en) * | 2020-05-12 | 2020-10-13 | 中铁第四勘察设计院集团有限公司 | Deep well aeration process optimization method and device for sewage treatment plant based on ASM2D model |
| WO2022256682A1 (en) * | 2021-06-03 | 2022-12-08 | The Texas A&M University System | Systems and methods for determining constitutive parameters of subject materials |
| CN114171127A (en) * | 2021-12-07 | 2022-03-11 | 上海交通大学四川研究院 | Method for simulating supersonic combustion flow by constructing self-adaptive reaction mechanism |
Non-Patent Citations (1)
| Title |
|---|
| 吴坤;范学军: "超声速燃烧数值模拟中复杂化学反应的建模方法", 《空气动力学学报》, 14 April 2020 (2020-04-14), pages 552 - 576 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118260975B (en) | 2024-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bove et al. | Modeling solid oxide fuel cell operation: Approaches, techniques and results | |
| US20210028472A1 (en) | Similar Principle Analysis Method of Input and Output Characteristics for Fuel Cell | |
| CN109767815B (en) | Method for simplifying combustion reaction mechanism based on rate uncertainty | |
| CN107301583A (en) | It is a kind of that method is recommended based on user preference and the cold start-up trusted | |
| Verduzco et al. | An active learning approach for the design of doped LLZO ceramic garnets for battery applications | |
| CN107228926A (en) | The explosive logic network analysis method for reliability assessed based on accelerated aging | |
| CN118260975B (en) | Engine combustion chamber numerical simulation method based on self-adaptive pre-zoning mechanism | |
| Wang et al. | CrabNet for explainable deep learning in materials science: bridging the gap between academia and industry | |
| CN114171127B (en) | Method for constructing self-adaptive reaction mechanism to simulate supersonic combustion flow | |
| Choi et al. | Multi-variate optimization of polymer electrolyte membrane fuel cells in consideration of effects of GDL compression and intrusion | |
| Mert et al. | Multi‐objective optimization of a direct methanol fuel cell system using a genetic‐based algorithm | |
| Wei et al. | Research on PEMFC internal temperature predictions and thermal management strategy based on a Kalman algorithm | |
| Kumar et al. | Genetic algorithm‐based higher‐order model reduction of proton exchange membrane fuel cell | |
| CN110069856A (en) | CFD system and implementation method based on adaptive reaction mechanism | |
| CN113222216B (en) | Cold and hot electric load prediction method, device and system | |
| Yang et al. | Prediction on Discharging Properties of Nickel–Manganese Materials for High‐Performance Sodium‐Ion Batteries via Machine Learning Methods | |
| Li et al. | Multi-objective optimization design of a planar membrane humidifier based on NSGA-II and entropy weight TOPSIS | |
| CN115662542A (en) | Method for determining influence of W doping on thermoelectric properties of bismuth selenide oxide | |
| CN116364212A (en) | Screening optimization method and material for machine learning model acceleration fast ion conductor material | |
| Nakabayashi et al. | A process simulator based on hybrid model of physical model and Just-In-Time model | |
| He et al. | Exploring flux representations of complex kinetics networks | |
| FR3122767A1 (en) | Method for constructing a simulation model of a chemical reaction | |
| CN102339342A (en) | Quick materialization method of parameterized device unit | |
| WO2024014094A1 (en) | Characteristic prediction device, characteristic prediction method, and characteristic prediction program | |
| Zheng et al. | A Novel Data‐Driven Approach to Lithium‐ion Battery Dynamic Charge State Capture for New Energy Electric Vehicles |
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 |