CN114861364A - Intelligent sensing and suction regulation and control method for air inlet flow field of air-breathing engine - Google Patents
Intelligent sensing and suction regulation and control method for air inlet flow field of air-breathing engine Download PDFInfo
- Publication number
- CN114861364A CN114861364A CN202210596661.XA CN202210596661A CN114861364A CN 114861364 A CN114861364 A CN 114861364A CN 202210596661 A CN202210596661 A CN 202210596661A CN 114861364 A CN114861364 A CN 114861364A
- Authority
- CN
- China
- Prior art keywords
- suction
- flow field
- pumping
- air
- back pressure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
-
- 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
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
-
- 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
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Geometry (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Algebra (AREA)
- Computing Systems (AREA)
- Fluid Mechanics (AREA)
- Mathematical Physics (AREA)
- Feedback Control In General (AREA)
Abstract
The invention provides an intelligent sensing and suction regulation method for a flow field of an air inlet passage of an air-breathing engine, which performs intelligent suction regulation and control on a hypersonic air inlet passage according to a real-time sensed flow field state so as to achieve the aim of intelligently controlling the flow field state, and reduces the loss of excessive flow while widening the running boundary of the engine. And under the conditions of high efficiency and high precision, carrying out real-time adaptive control on the optimized pumping parameters by using an active disturbance rejection control system.
Description
Technical Field
The invention belongs to the technical field of intelligent sensing and suction regulation of a flow field of an air inlet passage of an air-breathing engine, and particularly relates to an intelligent sensing and suction regulation method of the flow field of the air inlet passage of the air-breathing engine.
Background
The hypersonic weapon using the air suction type engine as power has the outstanding characteristics of high flying speed, strong maneuverability and the like. The air intake passage is critical to the air-breathing engine, and the limited back pressure resistance of the air intake passage imposes a limit on the performance of the air-breathing engine. In only a few times of hypersonic flight tests which are historically carried out, a failure case of the flight test caused by flow instability appears for a plurality of times, which shows that people have unclear understanding of the complex flow phenomenon and mechanism in the engine and have insufficient control capability. Therefore, it is a great need to achieve high maneuvering safety and high efficiency operation of the engine over a wide mach number range.
At present, a suction flow control method is mostly adopted at home and abroad to carry out boundary layer suction control aiming at the problems of flow instability and narrow operation boundary of an air inlet passage of an air suction type engine. Although the backpressure resistance of the air inlet passage can be improved and the operation boundary of the engine can be widened, the lost flow can have adverse effects on the performance of the engine. Although the influence of different boundary layer suction parameters (such as suction back pressure, suction position, suction rate, use of suction holes or suction grooves and hole/groove area) on the engine flow field is developed, the influence is basically in the aspect of manually modifying control parameters, models and the like.
Therefore, it is urgently needed to develop an intelligent pumping regulation and control method, which reasonably adjusts the opening and closing of the pumping groove and the pumping flow rate by combining the working state of the air inlet channel according to the real-time perceived flow field state, so as to find the goals of optimizing the optimal control parameters and performance, further widen the operation boundary of the engine, reduce the loss of excessive flow rate, and provide support for improving the performance of the engine.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide an intelligent sensing and suction regulation method for a flow field of an air inlet passage of an air-breathing engine, so as to realize self-adaptive suction control of the hypersonic-speed air inlet passage.
In order to achieve the purpose, the technical scheme of the invention is as follows:
an intelligent sensing and suction regulation and control method for an air inlet channel flow field of an air-breathing engine comprises the following steps:
s1, analyzing the influence of wall pressure on flow field reconstruction aiming at an air inlet channel of the air-breathing engine; analyzing the influence of the main pumping parameters on the counter-pressure resistance;
s2, acquiring a flow field reconstruction data set through numerical simulation and ground experiments under different inflow conditions and working conditions of the scramjet; obtaining sample points of a pumping parameter design space by using a Latin hypercube sampling method, and obtaining a pumping data set through a ground experiment;
s3, preprocessing the flow field reconstruction data set and the suction data set;
s4, building a depth network model to form a flow field reconstruction agent model and a suction agent model;
s5, searching a design variable which meets the requirements of the optimal front edge position, stability margin and back pressure resistance of the shock wave string by using an immune heuristic multi-objective optimization algorithm and combining a flow field reconstruction agent model and a suction agent model;
s6, performing real-time anti-interference regulation and control on pumping parameters by using an active disturbance rejection control algorithm according to the global optimal variable obtained by the immune heuristic multi-target algorithm;
and S7, performing intelligent pumping regulation according to the flow field state sensed in real time.
Preferably, step S1 is specifically:
analyzing the influence of wall pressure on flow field reconstruction aiming at an air inlet channel of an air-breathing engine; analyzing the influence of main pumping parameters on the anti-back pressure capability, observing the front edge position of the shock wave string and the stability margin of the shock wave string as a result of flow field reconstruction, wherein the main pumping parameters comprise pumping back pressure, pumping position, pumping rate and the number of used pumping holes, and further obtaining the mathematical relationship between the wall pressure and the flow field reconstruction and the mathematical relationship between the pumping parameters and the anti-back pressure capability.
Preferably, step S2 is specifically:
acquiring a flow field reconstruction data set through numerical simulation and ground experiments under different incoming flow conditions and working conditions of the air-breathing engine; carrying out parameterization processing on 4 design variables of suction back pressure, suction position, suction rate and the number of used suction holes, determining the upper bound and the lower bound of each design variable, acquiring sample points of a design space by combining a Latin hypercube sampling method, and acquiring a suction data set through a ground experiment.
As a preferable mode, the step S2 of obtaining the sample points of the design space by combining the latin hypercube sampling method specifically includes:
(1) firstly, determining the number M of samples to be acquired;
(2) respectively dividing intervals of 4 design variables of suction back pressure, suction position, suction rate and the number of used suction holes into M sections;
(3) randomly drawing a value in each of the M sections;
(4) randomly combining the values extracted by different design variables;
(5) and obtaining M groups of sampled samples.
And taking the obtained M groups of samples as input conditions of a ground test, obtaining the magnitude of the anti-back pressure capacity through the ground test, verifying and correcting a Computational Fluid Dynamics (CFD) numerical simulation result, expanding the data sample size through the Computational Fluid Dynamics (CFD) numerical simulation, and constructing a data set with design variables corresponding to the combustion performance one by one.
Preferably, step S3 is specifically:
and analyzing and preprocessing the flow field reconstruction data set and the suction data set, screening and eliminating completely worthless data to ensure the quality of the basic data set, wherein the completely worthless data comprises abnormal data, redundant data and data outside a design variable range.
Preferably, step S4 is specifically:
constructing an agent model based on a depth network, and constructing a mathematical model of the wall surface pressure of the air inlet passage and the front edge position and the stability margin of the shock wave string under different incoming flows and working conditions by combining a preprocessed data set to form an agent model for reconstructing a flow field; meanwhile, constructing a proxy model of suction by using suction backpressure, suction position, suction rate, 4 design variables of the number of used suction holes and the anti-backpressure capacity;
for the flow field reconstruction proxy model, the input of the depth network is the wall surface pressure of an air inlet channel, and the output is the shock wave string front edge position and the stability margin of a flow field image; for the proxy model of pumping, the input of the depth network is 4 design variables such as pumping back pressure, pumping position, pumping rate and the number of used pumping holes, and the output is the anti-back pressure capacity.
Preferably, step S5 is specifically:
under different inflow conditions and different working conditions of the air-breathing engine, a flow field reconstruction agent model and a suction agent model established by a depth network are used as fitness functions, and an immune heuristic multi-objective optimization algorithm is utilized to find suction backpressure, suction position, suction rate and the number of used suction holes which enable the front edge position, stability margin and back pressure resistance of a shock wave string to be optimal, namely the optimal values of 4 design variables are found.
Preferably, in step S5, the process of finding the optimal pumping design variable by the immune heuristic multi-objective optimization algorithm includes:
(1) antigen recognition, namely understanding a problem to be optimized, performing feasibility analysis on the problem, extracting prior knowledge, constructing a proper affinity function, and formulating various constraint conditions;
(2) an initial population of antibodies, representing feasible solutions to the problem as antibodies in a solution space by encoding, randomly generating an initial population within the solution space;
(3) performing affinity evaluation on each feasible solution in the population;
(4) judging whether an algorithm termination condition is met; if the condition is met, the algorithm optimizing process is stopped, and a calculation result is output; otherwise, continuing the optimization operation;
(5) calculating the concentration and the excitation degree of the antibody;
(6) performing immune treatment including immune selection, cloning, mutation and clone inhibition;
immune selection: selecting high-quality antibodies according to the calculation results of the affinity and the concentration of the antibodies in the population, and activating the high-quality antibodies;
cloning: cloning and copying the activated antibody to obtain a plurality of copies;
mutation: carrying out mutation operation on the copy obtained by cloning to enable the copy to generate affinity mutation;
cloning inhibition: reselecting the compiling result, inhibiting the antibody with low affinity, and keeping the variation result with high affinity;
(7) and (4) refreshing the population, replacing the antibody with low excitation degree in the population with the randomly generated new antibody to form a new generation antibody, and turning to the step (3).
Preferably, step S6 is specifically:
according to an immune heuristic multi-target algorithm, determining optimal values of a group of controllable variables, namely the corresponding suction backpressure, suction position, suction rate and the number of used suction holes under the condition that the position of the front edge of a shock wave string, the stability margin and the back pressure resistance are optimal; and (3) an active disturbance rejection control system is built, and real-time adaptive control is carried out on suction backpressure and suction rate, so that adaptive control with optimal suction performance is realized.
Preferably, in step S6, the active disturbance rejection control system specifically includes:
an active disturbance rejection control system (ADRC) is used for controlling easily interfered and unstable suction variables in the flight process, namely, the real-time control of suction backpressure and suction rate is completed, and a motor for controlling suction is easily interfered in the flight process, so that the variables directly controlled by the motor are easily interfered and unstable variables;
the active disturbance rejection control system consists of a tracking differentiator, an extended state observer and nonlinear error feedback control; the tracking differentiator is used for tracking and differentiating the input signal of the system to obtain a stable system input signal; the extended state observer is used for estimating the real-time state of the pumping parameters and the total disturbance of the system; the nonlinear error feedback control is used for compensating the control rate according to the stable system input signal of the system, the estimated real-time state and the total disturbance, and generating the final control quantity of the pumping parameters.
Preferably, step S7 is specifically:
the real-time intelligent sensing of the flow field is completed by a flow field reconstruction and suction decision system, and the flow field reconstruction and suction decision system comprises four parts of data acquisition and transmission, shock wave string front edge position detection, suction groove opening and closing judgment and suction decision;
according to different incoming flow disturbances and back pressure disturbances, combining the pressure data of the wall surface of the air inlet channel, and performing data acquisition and transmission by using a data acquisition and transmission platform; combining the acquired data with a depth network to obtain a real-time flow field image, detecting the position of the front edge of a real-time laser string and the stability margin, and simultaneously calculating the anti-back pressure capacity; taking the detected front edge position of the shock wave string, the stability margin and the back pressure resistance as conditions for judging whether to suck or not; if the suction condition is met, the suction groove is opened, and a suction decision is made; otherwise, the suction groove is closed;
in the pumping process, a flow field reconstruction agent model and a pumping agent model are used for respectively carrying out flow field reconstruction and pumping effect calculation, an immune heuristic multi-target optimization algorithm is combined to find pumping parameters which enable the front edge position, stability margin and anti-back pressure capability of a shock wave string to be optimal under the current condition, the obtained optimal parameters are transmitted to a pumping control system, the flow field state is further intelligently sensed, finally, a closed loop is formed in the whole pumping process, and the effects of real-time flow field sensing and intelligent pumping are achieved;
the optimized suction backpressure and suction rate are used as expectations of an ADRC control system, interference is filtered by the ADRC control system, suction parameters are stabilized, and intelligent suction energy can always keep the best performance under different conditions; the remaining two aspiration parameters, aspiration location and number of aspiration holes used, are used directly to make aspiration decisions.
The invention has the beneficial effects that: the invention provides an intelligent sensing and suction regulation method for a flow field of an air inlet passage of an air-breathing engine through the air-breathing engine, which performs intelligent suction regulation and control on a hypersonic air inlet passage according to a real-time sensed flow field state so as to achieve the aim of intelligently controlling the flow field state, and reduces the loss of excessive flow while widening the operation boundary of the engine. Under the conditions of high efficiency and high precision, the active disturbance rejection control system is utilized to carry out real-time active disturbance rejection control on the optimized pumping parameters, and pumping is subjected to self-adaptive adjustment under the conditions of complex nonlinear systems, uncertain external environments and the like.
Drawings
FIG. 1 is a schematic flow chart of an intelligent sensing and pumping regulation method for hypersonic inlet channel flow field information according to the invention;
FIG. 2 is a schematic structural diagram of an intelligent hypersonic inlet channel flow field information sensing and pumping regulation system of the present invention;
FIG. 3 is a schematic diagram of a convolutional neural network structure according to the present invention;
FIG. 4 is a schematic diagram of a fully-connected network architecture according to the present invention;
FIG. 5 is a flow chart diagram of an immune heuristic multi-objective optimization algorithm of the present invention;
FIG. 6 is a schematic diagram of a two-dimensional structure for hypersonic inlet pumping according to the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention.
Example 1
The embodiment provides an intelligent sensing and suction regulation method for an air inlet flow field of an air-breathing engine, which comprises the following steps as shown in fig. 1:
s1, analyzing the influence of wall pressure on flow field reconstruction aiming at an air inlet channel of the air-breathing engine; analyzing the influence of the main pumping parameters on the counter-pressure resistance;
s2, acquiring a flow field reconstruction data set through numerical simulation and ground experiments under different inflow conditions and working conditions of the scramjet; obtaining sample points of a pumping parameter design space by using a Latin hypercube sampling method, and obtaining a pumping data set through a ground experiment;
s3, preprocessing the flow field reconstruction data set and the suction data set;
s4, building a depth network model to form a flow field reconstruction agent model and a suction agent model;
s5, searching a design variable which meets the requirements of the optimal front edge position, stability margin and back pressure resistance of the shock wave string by using an immune heuristic multi-objective optimization algorithm and combining a flow field reconstruction agent model and a suction agent model;
s6, performing real-time anti-interference regulation and control on pumping parameters by using an active disturbance rejection control algorithm according to the global optimal variable obtained by the immune heuristic multi-target algorithm;
and S7, performing intelligent pumping regulation according to the flow field state sensed in real time.
Example 2:
the embodiment provides an intelligent sensing and suction regulation method for an air inlet flow field of an air-breathing engine, and specifically includes the following steps S1-S7 as shown in fig. 1 and fig. 2:
s1, analyzing the influence of wall pressure on flow field reconstruction aiming at an air inlet channel of the air-breathing engine; analyzing the influence of the main pumping parameters on the counter-pressure resistance;
the specific implementation manner of step S1 is:
analyzing the influence of wall pressure on flow field reconstruction aiming at an air inlet channel of an air-breathing engine; analyzing the influence of main pumping parameters on the anti-back pressure capability, observing the front edge position of the shock wave string and the stability margin of the shock wave string as a result of flow field reconstruction, wherein the main pumping parameters comprise pumping back pressure, pumping position, pumping rate and the number of used pumping holes, and further obtaining the mathematical relationship between the wall pressure and the flow field reconstruction and the mathematical relationship between the pumping parameters and the anti-back pressure capability.
S2, acquiring a flow field reconstruction data set through numerical simulation and ground experiments under different inflow conditions and working conditions of the air-breathing engine; obtaining sample points of a pumping parameter design space by using a Latin hypercube sampling method, and obtaining a pumping data set through a ground experiment;
the specific implementation manner of step S2 is:
acquiring a flow field reconstruction data set through numerical simulation and ground experiments under different incoming flow conditions and working conditions of the air-breathing engine; carrying out parameterization processing on 4 design variables of suction back pressure, suction position, suction rate and the number of used suction holes, determining the upper bound and the lower bound of each design variable, acquiring sample points of a design space by combining a Latin hypercube sampling method, and acquiring a suction data set through a ground experiment.
As a preferable mode, the step S2 of obtaining the sample points of the design space by combining the latin hypercube sampling method specifically includes:
(1) firstly, determining the number M of samples to be acquired;
(2) respectively dividing intervals of 4 design variables of suction back pressure, suction position, suction rate and the number of used suction holes into M sections;
(3) randomly drawing a value in each of the M sections;
(4) randomly combining the values extracted by different design variables;
(5) and obtaining M groups of sampled samples.
And taking the obtained M groups of samples as input conditions of a ground test, obtaining the magnitude of the anti-back pressure capacity through the ground test, verifying and correcting a Computational Fluid Dynamics (CFD) numerical simulation result, expanding the data sample size through the Computational Fluid Dynamics (CFD) numerical simulation, and constructing a data set with design variables corresponding to the combustion performance one by one.
S3, preprocessing the flow field reconstruction data set and the suction data set;
the specific implementation manner of step S3 is:
and analyzing and preprocessing the flow field reconstruction data set and the suction data set, screening and eliminating completely worthless data to ensure the quality of the basic data set, wherein the completely worthless data comprises abnormal data, redundant data and data outside a design variable range.
S4, building a depth network model to form a flow field reconstruction agent model and a suction agent model;
the specific implementation manner of step S4 is:
constructing an agent model based on a depth network, and constructing a mathematical model of the wall surface pressure of the air inlet passage and the front edge position and the stability margin of the shock wave string under different incoming flows and working conditions by combining a preprocessed data set to form an agent model for reconstructing a flow field; meanwhile, a proxy model of suction is constructed by using suction back pressure, suction position, suction rate and the size of anti-back pressure capacity by using 4 design variables of the number of suction holes.
Deep networks are a broad concept in the sense that Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN) and generative countermeasure networks (GAN) all fall into the category of deep networks. The invention takes CNN as an example, and the concrete mode is as follows:
the CNN includes a convolution layer, a pooling layer, and a full-link layer, as shown in fig. 3, the CNN can not only effectively reduce a picture with a large amount of data into a small amount of data, but also effectively retain picture characteristics, and conforms to the principle of picture processing. The convolution layer is composed of a group of convolution kernels, and features of input data are extracted after convolution operation is carried out on the input data and the convolution kernels. Multiple convolutions are generally required to extract the global features. The relationship between the length and the width of the data before and after the convolution operation is shown as follows:
wherein, W 1 And W 2 Respectively the data widths before and after the convolution operation; f is the width of the convolution kernel; s is the step length of convolution operation; p is the fill size. Meanwhile, the width of the data before and after the convolution operation is as follows:
wherein H 1 And H 2 The heights of the data before and after the convolution operation are respectively.
In the pooling layer, after the convolutional layer is usually arranged, the feature dimension output by the convolutional layer is reduced by performing local operation on input data, and the complexity of subsequent network training is reduced. The pooling operation downsamples the input feature map to obtain the most prominent information. The pooling layer generally employs an average pooling:
y pool ＝avg(x[i,j]),i∈[1,l],j∈[1,l]
where avg (·) is the averaging operation, x is the input matrix, and l is the pooling window size.
The fully connected layer performs the mapping of the features to the output layer and after connecting to the output layer of the network, the classification or estimation is completed.
Preferably, in step S4, as shown in fig. 4, the full connection process is:
the number of the layers of the full-connection network is K, and the number of neurons in each layer is m in sequence 0 ,m 1 ,m 2 ,…,m K . For the k-th layer of a fully connected network:
wherein the content of the first and second substances,
the ith neuron of the kth layer; m is k The number of neurons in the k-th layer;
the weight of the ith row and the jth column of the kth layer weight matrix is obtained;
is the bias of the k-th layer; w (k) Is a weight matrix of the k layer; y is (k-1) Is the output vector of the k-1 layer and the input vector of the k layer; y is (k) Is the output vector of the k layer; here net (k) Each element in the k-th layer represents the weighted sum of the k-th layer input vector multiplied by the weight matrix and the offset vector; f. of (k) As a function of activation of layer k neurons.
For the flow field reconstruction proxy model, the input of the depth network is the wall surface pressure of an air inlet channel, and the output is the shock wave string front edge position and the stability margin of a flow field image; for the proxy model of pumping, the input of the depth network is 4 design variables such as pumping back pressure, pumping position, pumping rate and the number of used pumping holes, and the output is the anti-back pressure capacity.
S5, searching a design variable which meets the requirements of the optimal front edge position, stability margin and back pressure resistance of the shock wave string by using an immune heuristic multi-objective optimization algorithm and combining a flow field reconstruction agent model and a suction agent model;
the specific implementation manner of step S5 is:
under different inflow conditions and different working conditions of the air-breathing engine, as shown in fig. 5, a flow field reconstruction agent model and a suction agent model established by a depth network are used as fitness functions, and an immune heuristic multi-objective optimization algorithm is utilized to find suction backpressure, suction position, suction rate and the number of used suction holes which enable the front edge position, stability margin and anti-back pressure capacity of a shock wave string to be optimal, namely, the optimal values of 4 design variables are found.
Preferably, in step S5, the process of finding the optimal pumping design variable by the immune heuristic multi-objective optimization algorithm includes:
(1) antigen recognition, namely understanding a problem to be optimized, performing feasibility analysis on the problem, extracting prior knowledge, constructing a proper affinity function, and formulating various constraint conditions.
(2) An initial population of antibodies is randomly generated within the solution space by encoding antibodies that represent feasible solutions to the problem as the solution space.
(3) Affinity evaluation was performed for each feasible solution in the population.
(4) Judging whether an algorithm termination condition is met; if the condition is met, the algorithm optimizing process is stopped, and a calculation result is output; otherwise, continuing the optimizing operation.
(5) Antibody concentration and degree of excitation were calculated.
(6) Immunological processing including immunoselection, cloning, mutation, and clonal suppression is performed.
And (3) immune selection: selecting high-quality antibodies according to the calculation results of the affinity and the concentration of the antibodies in the population, and activating the high-quality antibodies;
cloning: cloning and copying the activated antibody to obtain a plurality of copies;
mutation: carrying out mutation operation on the copy obtained by cloning to enable the copy to generate affinity mutation;
clone inhibition: and (4) reselecting the compiling result, inhibiting the antibody with low affinity, and keeping the variation result with high affinity.
(7) And (4) refreshing the population, replacing the antibody with low excitation degree in the population with the randomly generated new antibody to form a new generation of antibody, and turning to the step (3).
S6, performing real-time anti-interference regulation and control on pumping parameters by using an active disturbance rejection control algorithm according to the global optimal variable obtained by the immune heuristic multi-target algorithm;
the specific implementation manner of step S6 is:
and determining the optimal values of a group of controllable variables according to an immune heuristic multi-objective algorithm, namely the corresponding suction backpressure, suction position, suction rate and the number of used suction holes under the condition that the position of the front edge of the shock wave string, the stability margin and the back pressure resistance capability are optimal. And (3) an active disturbance rejection control system is built, and real-time adaptive control is carried out on suction backpressure and suction rate, so that adaptive control with optimal suction performance is realized.
Preferably, in step S6, the operation and configuration of the active disturbance rejection control system specifically refer to:
an active disturbance rejection control system (ADRC) is used for controlling the easily disturbed and unstable suction variables in the flight process, namely, the real-time control of suction backpressure and suction rate is completed, and the motor for controlling suction is easily disturbed in the flight process, so the variables directly controlled by the motor are easily disturbed and unstable variables;
the active disturbance rejection control system consists of a tracking differentiator, an extended state observer and nonlinear error feedback control; the tracking differentiator is used for tracking and differentiating the input signal of the system to obtain a stable system input signal; the extended state observer is used for estimating the real-time state of the pumping parameters and the total disturbance of the system; the nonlinear error feedback control is used for compensating the control rate according to the stable system input signal of the system, the estimated real-time state and the total disturbance, and generating the final control quantity of the pumping parameters.
At the suction rate v during suction 0 For example, the suction back pressure is also controlled by the method and procedure of the suction rate. The input of the ADRC system is the output v under the optimized system 0 The output of the ADRC system is the final control U of the aspiration rate.
The input to the differential controller is the desired suction rate v 0 Output as an input signal v 0 First order tracking differential signal v 1 And a second order tracking differential signal v 2 . The differential controller is as follows:
in the formula, r, v 0 Parameters to be adjusted; h is the operation step length; k is the number of sampling moments; v. of 1 For the input signal v at time k 0 Tracking input signal of v 2 (k) Is v 1 (k) The first order differential signal of (1); sign () is a sign function; fhan () is the fastest control synthesis function; a is 0 Is the boundary layer thickness, y, s y And s a Are all intermediate parameters.
The inputs to the extended state observer are: actual pumping rate y, control quantity U and coefficient b of pumping control system 0 The product of (a) and (b). The outputs are respectively the observed suction rate z 1 Observing the rate of change of aspiration rate z 2 And observing the total disturbance z 3 . Wherein the observed aspiration rate is considered to be the actual aspiration rate; observing the rate of change of aspiration rate may be considered the actual rate of change of aspiration rate; the observed total disturbance is the total disturbance inside and outside the system, which is divided by b 0 Then, the state error feedback control law output u is used 0 Subtracting to obtain the final control quantity U of the pumping rate of the system:
U(k)＝u 0 (k)-z 3 (k)/b 0
in the formula, k is the number of sampling moments; u (k) is the final control output quantity of the ADRC subsystem at the moment k; z is a radical of 3 (k) Is the total disturbance observed at time k; b 0 To determine the compensation factor for compensating the strength.
The extended state observer is:
in the formula, k is the number of sampling moments; z is a radical of 1 (k) The observed aspiration rate at time k; z is a radical of 2 (k) Is the observed pumping rate change rate at time k; z is a radical of 3 (k) The total disturbance estimation value of the internal and external disturbances in the pumping at the moment k is obtained; z is a radical of 1 (k+1)、z 2 (k+1)、z 3 (k +1) respectively representing the observed pumping rate, the observed pumping rate change rate and the total disturbance estimation value of internal and external disturbance during pumping at the k +1 moment; a is 1 ,a 2 ,a 3 ,β 01 ,β 02 And beta 03 Are all adjustment parameters; δ is the zone length of the linear segment.
The nonlinear error feedback input is the aspiration rate error e 1 (tracking aspiration Rate v) 1 Observation of the aspiration rate z 1 ) Error of rate of change e 2 (tracking the Rate of Change of suction v 2 Observation of the rate of change of aspiration rate z 2 ) Output the control quantity U 0 The nonlinear error feedback is processed as follows:
in the formula u 0 The error feedback control quantity is used; beta is a 1 And beta 2 Are all adjustable parameters; e.g. of the type 1 For systematic suction rate error, e 2 Is the system pumping rate error.
And S7, performing intelligent pumping regulation according to the flow field state sensed in real time.
The specific implementation manner of step S7 is:
real-time intelligent sensing of the flow field is accomplished by a flow field reconstruction and aspiration decision system, as shown in fig. 2. The flow field reconstruction and suction decision system comprises four parts, namely high-efficiency and high-fidelity data acquisition and transmission, shock wave string front edge position detection, suction groove opening and closing judgment and suction decision.
According to different incoming flow disturbances and back pressure disturbances, combining the pressure data of the wall surface of the air inlet channel, and utilizing a high-efficiency high-fidelity platform to acquire and transmit data; combining the acquired data with a depth network to obtain a real-time flow field image, detecting the position of the front edge of a real-time laser string and the stability margin, and simultaneously calculating the anti-back pressure capacity; taking the detected front edge position of the shock wave string, the stability margin and the back pressure resistance as conditions for judging whether to suck or not; if the suction condition is met, the suction groove is opened, and a suction decision is made; otherwise, the suction groove is closed;
in the pumping process, a flow field reconstruction agent model and a pumping agent model are used for respectively carrying out flow field reconstruction and pumping effect calculation, an immune heuristic multi-target optimization algorithm is combined to find pumping parameters which enable the front edge position, stability margin and anti-back pressure capability of a shock wave string to be optimal under the current condition, the obtained optimal parameters are transmitted to a pumping control system, the flow field state is further intelligently sensed, finally, a closed loop is formed in the whole pumping process, and the effects of real-time flow field sensing and intelligent pumping are achieved;
the optimized suction backpressure and suction rate are used as expectations of an active disturbance rejection control system (ADRC), interference is filtered by the ADRC, suction parameters are stabilized, and the intelligent suction energy can always keep the best performance under different conditions; the remaining two aspiration parameters, aspiration location and number of aspiration holes used, are used directly to make aspiration decisions.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (11)
1. An intelligent sensing and suction regulation and control method for an air inlet channel flow field of an air-breathing engine is characterized by comprising the following steps:
s1, analyzing the influence of wall pressure on flow field reconstruction aiming at an air inlet channel of the air-breathing engine; analyzing the influence of the main pumping parameters on the counter-pressure resistance;
s2, acquiring a flow field reconstruction data set through numerical simulation and ground experiments under different inflow conditions and working conditions of the air-breathing engine; obtaining sample points of a pumping parameter design space by using a Latin hypercube sampling method, and obtaining a pumping data set through a ground experiment;
s3, preprocessing the flow field reconstruction data set and the suction data set;
s4, building a depth network model to form a flow field reconstruction agent model and a suction agent model;
s5, searching a design variable which meets the requirements of the optimal front edge position, stability margin and back pressure resistance of the shock wave string by using an immune heuristic multi-objective optimization algorithm and combining a flow field reconstruction agent model and a suction agent model;
s6, performing real-time anti-interference regulation and control on pumping parameters by using an active disturbance rejection control algorithm according to the global optimal variable obtained by the immune heuristic multi-target algorithm;
and S7, performing intelligent pumping regulation according to the flow field state sensed in real time.
2. The method for intelligently sensing and regulating suction of the flow field of the air intake passage of the air-breathing engine according to claim 1, wherein the step S1 is specifically as follows:
analyzing the influence of wall pressure on flow field reconstruction aiming at an air inlet channel of an air-breathing engine; analyzing the influence of main pumping parameters on the anti-back pressure capability, observing the front edge position of the shock wave string and the stability margin of the shock wave string as a result of flow field reconstruction, wherein the main pumping parameters comprise pumping back pressure, pumping position, pumping rate and the number of used pumping holes, and further obtaining the mathematical relationship between the wall pressure and the flow field reconstruction and the mathematical relationship between the pumping parameters and the anti-back pressure capability.
3. The method for intelligently sensing and suction regulating of the inlet flow field of the air-breathing engine according to claim 1, wherein the step S2 is specifically as follows:
acquiring a flow field reconstruction data set through numerical simulation and ground experiments under different incoming flow conditions and working conditions of the air-breathing engine; carrying out parameterization processing on 4 design variables of suction back pressure, suction position, suction rate and the number of used suction holes, determining the upper bound and the lower bound of each design variable, acquiring sample points of a design space by combining a Latin hypercube sampling method, and acquiring a suction data set through a ground experiment.
4. The method for intelligently sensing and suction regulating of the intake passage flow field of the air-breathing engine according to claim 1, wherein the step S2 of obtaining the sample points of the design space in combination with the latin hypercube sampling method specifically comprises the steps of:
(1) firstly, determining the number M of samples to be acquired;
(2) respectively dividing intervals of 4 design variables of suction back pressure, suction position, suction rate and the number of used suction holes into M sections;
(3) randomly drawing a value in each of the M sections;
(4) randomly combining the values extracted by different design variables;
(5) obtaining M groups of sampled samples;
and taking the obtained M groups of samples as input conditions of a ground test, obtaining the magnitude of the anti-back pressure capacity through the ground test, verifying and correcting a Computational Fluid Dynamics (CFD) numerical simulation result, expanding the data sample size through the Computational Fluid Dynamics (CFD) numerical simulation, and constructing a data set with design variables corresponding to the combustion performance one by one.
5. The method for intelligently sensing and suction regulating of the inlet flow field of the air-breathing engine according to claim 1, wherein the step S3 is specifically as follows:
and analyzing and preprocessing the flow field reconstruction data set and the suction data set, screening and eliminating completely worthless data to ensure the quality of the basic data set, wherein the completely worthless data comprises abnormal data, redundant data and data outside a design variable range.
6. The method for intelligently sensing and suction regulating of the inlet flow field of the air-breathing engine according to claim 1, wherein the step S4 is specifically as follows:
constructing an agent model based on a depth network, and constructing a mathematical model of the wall surface pressure of the air inlet passage and the front edge position and the stability margin of the shock wave string under different incoming flows and working conditions by combining a preprocessed data set to form an agent model for reconstructing a flow field; meanwhile, constructing a proxy model of suction by using suction backpressure, suction position, suction rate, 4 design variables of the number of used suction holes and the anti-backpressure capacity;
for the flow field reconstruction proxy model, the input of the depth network is the wall surface pressure of an air inlet channel, and the output is the shock wave string front edge position and the stability margin of a flow field image; for the proxy model of pumping, the input of the depth network is 4 design variables such as pumping back pressure, pumping position, pumping rate and the number of used pumping holes, and the output is the anti-back pressure capacity.
7. The method for intelligently sensing and suction regulating of the inlet flow field of the air-breathing engine according to claim 1, wherein the step S5 is specifically as follows:
under different inflow conditions and different working conditions of the air-breathing engine, a flow field reconstruction agent model and a suction agent model established by a depth network are used as fitness functions, and an immune heuristic multi-objective optimization algorithm is utilized to find suction backpressure, suction position, suction rate and the number of used suction holes which enable the front edge position, stability margin and back pressure resistance of a shock wave string to be optimal, namely the optimal values of 4 design variables are found.
8. The method for intelligently sensing and regulating the flow field of the air intake passage of the air-breathing engine as claimed in claim 1, wherein in the step S5, the process of searching for the optimal pumping design variable by the immune heuristic multi-objective optimization algorithm comprises the following steps:
(1) antigen recognition, namely understanding a problem to be optimized, performing feasibility analysis on the problem, extracting prior knowledge, constructing a proper affinity function, and formulating various constraint conditions;
(2) an initial population of antibodies, representing feasible solutions to the problem as antibodies in a solution space by encoding, randomly generating an initial population within the solution space;
(3) performing affinity evaluation on each feasible solution in the population;
(4) judging whether an algorithm termination condition is met; if the condition is met, the algorithm optimizing process is stopped, and a calculation result is output; otherwise, continuing the optimization operation;
(5) calculating the concentration and the excitation degree of the antibody;
(6) performing immune treatment including immune selection, cloning, mutation and clone inhibition;
immune selection: selecting high-quality antibodies according to the calculation results of the affinity and the concentration of the antibodies in the population, and activating the high-quality antibodies;
cloning: cloning and copying the activated antibody to obtain a plurality of copies;
mutation: carrying out mutation operation on the copy obtained by cloning to enable the copy to generate affinity mutation;
clone inhibition: reselecting the compiling result, inhibiting the antibody with low affinity, and keeping the variation result with high affinity;
(7) and (4) refreshing the population, replacing the antibody with low excitation degree in the population with the randomly generated new antibody to form a new generation antibody, and turning to the step (3).
9. The method for intelligently sensing and suction regulating of the inlet flow field of the air-breathing engine according to claim 1, wherein the step S6 is specifically as follows:
according to an immune heuristic multi-target algorithm, determining optimal values of a group of controllable variables, namely the corresponding suction backpressure, suction position, suction rate and the number of used suction holes under the condition that the position of the front edge of a shock wave string, the stability margin and the back pressure resistance are optimal; and (3) an active disturbance rejection control system is built, and real-time adaptive control is carried out on suction backpressure and suction rate, so that adaptive control with optimal suction performance is realized.
10. The method for intelligently sensing and suction regulating of an air intake passage flow field of an air-breathing engine according to claim 9, wherein in step S6, the active disturbance rejection control system specifically comprises:
an active disturbance rejection control system (ADRC) is used for controlling the easily disturbed and unstable suction variables in the flight process, namely, the real-time control of suction backpressure and suction rate is completed, and the motor for controlling suction is easily disturbed in the flight process, so the variables directly controlled by the motor are easily disturbed and unstable variables;
the active disturbance rejection control system consists of a tracking differentiator, an extended state observer and nonlinear error feedback control; the tracking differentiator is used for tracking and differentiating the input signal of the system to obtain a stable system input signal; the extended state observer is used for estimating the real-time state of the pumping parameters and the total disturbance of the system; the nonlinear error feedback control is used for compensating the control rate according to the stable system input signal of the system, the estimated real-time state and the total disturbance, and generating the final control quantity of the pumping parameters.
11. The method for intelligently sensing and suction regulating of the inlet flow field of the air-breathing engine according to claim 1, wherein the step S7 is specifically as follows:
the real-time intelligent sensing of the flow field is completed by a flow field reconstruction and suction decision system, and the flow field reconstruction and suction decision system comprises four parts of data acquisition and transmission, shock wave string front edge position detection, suction groove opening and closing judgment and suction decision;
according to different incoming flow disturbances and back pressure disturbances, combining the pressure data of the wall surface of the air inlet channel, and performing data acquisition and transmission by using a data acquisition and transmission platform; combining the acquired data with a depth network to obtain a real-time flow field image, detecting the position of the front edge of a real-time laser string and the stability margin, and simultaneously calculating the anti-back pressure capacity; taking the detected front edge position of the shock wave string, the stability margin and the back pressure resistance as conditions for judging whether to suck or not; if the suction condition is met, the suction groove is opened, and a suction decision is made; otherwise, the suction groove is closed;
in the pumping process, a flow field reconstruction agent model and a pumping agent model are used for respectively carrying out flow field reconstruction and pumping effect calculation, an immune heuristic multi-target optimization algorithm is combined to find pumping parameters which enable the front edge position, stability margin and anti-back pressure capability of a shock wave string to be optimal under the current condition, the obtained optimal parameters are transmitted to a pumping control system, the flow field state is further intelligently sensed, finally, a closed loop is formed in the whole pumping process, and the effects of real-time flow field sensing and intelligent pumping are achieved;
the optimized suction backpressure and suction rate are used as expectations of an ADRC control system, interference is filtered by the ADRC control system, suction parameters are stabilized, and intelligent suction energy can always keep the best performance under different conditions; the remaining two aspiration parameters, aspiration location and number of aspiration holes used, are used directly to make aspiration decisions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210596661.XA CN114861364A (en) | 2022-05-30 | 2022-05-30 | Intelligent sensing and suction regulation and control method for air inlet flow field of air-breathing engine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210596661.XA CN114861364A (en) | 2022-05-30 | 2022-05-30 | Intelligent sensing and suction regulation and control method for air inlet flow field of air-breathing engine |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114861364A true CN114861364A (en) | 2022-08-05 |
Family
ID=82641700
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210596661.XA Pending CN114861364A (en) | 2022-05-30 | 2022-05-30 | Intelligent sensing and suction regulation and control method for air inlet flow field of air-breathing engine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114861364A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117436210A (en) * | 2023-12-18 | 2024-01-23 | 潍柴动力股份有限公司 | Combined design method and device for widening flow grooves and impellers |
-
2022
- 2022-05-30 CN CN202210596661.XA patent/CN114861364A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117436210A (en) * | 2023-12-18 | 2024-01-23 | 潍柴动力股份有限公司 | Combined design method and device for widening flow grooves and impellers |
| CN117436210B (en) * | 2023-12-18 | 2024-03-19 | 潍柴动力股份有限公司 | Combined design method and device for widening flow grooves and impellers |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109948841B (en) | Prediction method for residual oil distribution of water flooding oil field based on deep learning | |
| CN104200268B (en) | PSO (Particle Swarm Optimization) extremity learning machine based strip steel exit thickness predicting method | |
| CN106022954B (en) | Multiple BP neural network load prediction method based on grey correlation degree | |
| CN112083498B (en) | Multi-wave earthquake oil and gas reservoir prediction method based on deep neural network | |
| CN109343505A (en) | Gear method for predicting residual useful life based on shot and long term memory network | |
| CN111047094A (en) | Meter reading data anomaly analysis method based on deep learning algorithm | |
| CN108647583A (en) | A kind of face recognition algorithms training method based on multiple target study | |
| CN112733449A (en) | CNN well-seismic joint inversion method, CNN well-seismic joint inversion system, CNN well-seismic joint inversion storage medium, CNN well-seismic joint inversion equipment and CNN well-seismic joint inversion application | |
| Van Breusegem et al. | Adaptive neural models for on‐line prediction in fermentation | |
| CN111445010B (en) | Distribution network voltage trend early warning method based on evidence theory fusion quantum network | |
| CN109472397B (en) | Polymerization process parameter adjusting method based on viscosity change | |
| CN111260124A (en) | Chaos time sequence prediction method based on attention mechanism deep learning | |
| CN109740057B (en) | Knowledge extraction-based enhanced neural network and information recommendation method | |
| CN108550163A (en) | Moving target detecting method in a kind of complex background scene | |
| CN108596044B (en) | Pedestrian detection method based on deep convolutional neural network | |
| CN111624522A (en) | Ant colony optimization-based RBF neural network control transformer fault diagnosis method | |
| CN114861364A (en) | Intelligent sensing and suction regulation and control method for air inlet flow field of air-breathing engine | |
| CN110852440B (en) | Ocean front prediction method based on dynamic fuzzy neural network | |
| CN114254554A (en) | Long-time unstable well testing interpretation method based on LSTM model | |
| CN113627240A (en) | Unmanned aerial vehicle tree species identification method based on improved SSD learning model | |
| CN108681241A (en) | A kind of two-tank system discrimination method based on neural network | |
| Li et al. | Reservoir property prediction using the dynamic radial basis function network | |
| CN108446506B (en) | Uncertain system modeling method based on interval feedback neural network | |
| CN114815616A (en) | Intelligent regulation and control method and system for mode conversion of turbine stamping combined type engine | |
| CN109507989A (en) | A kind of method of unmanned plane sensor fault diagnosis |
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 |