CN112965382A - Nonlinear global sliding mode model-free control method based on neural network observer - Google Patents

Nonlinear global sliding mode model-free control method based on neural network observer Download PDF

Info

Publication number
CN112965382A
CN112965382A CN202110178100.3A CN202110178100A CN112965382A CN 112965382 A CN112965382 A CN 112965382A CN 202110178100 A CN202110178100 A CN 202110178100A CN 112965382 A CN112965382 A CN 112965382A
Authority
CN
China
Prior art keywords
neural network
model
sliding mode
control method
nonlinear
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110178100.3A
Other languages
Chinese (zh)
Other versions
CN112965382B (en
Inventor
张广明
吕筱东
高鹏
柏志青
扈凯
杨路
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Tech University
Original Assignee
Nanjing Tech University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing Tech University filed Critical Nanjing Tech University
Priority to CN202110178100.3A priority Critical patent/CN112965382B/en
Publication of CN112965382A publication Critical patent/CN112965382A/en
Application granted granted Critical
Publication of CN112965382B publication Critical patent/CN112965382B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a nonlinear global sliding mode model-free control method based on a neural network observer, which comprises the steps of establishing a mathematical model of a hypersonic aircraft aerodynamic heating ground simulation system and a super-local model without model control according to an energy conservation law; constructing a radial cubic b-spline-based adaptive neural network by utilizing a cubic b-spline basis function, and predicting unknown disturbance of the hypersonic aircraft aerodynamic heat ground simulation system; based on high-frequency switching buffeting in a nonlinear global sliding mode surface weakening approaching state; and establishing a nonlinear equivalent control rate and a nonlinear approach rate according to the sliding mode accessibility condition to obtain a nonlinear global sliding mode control rate. The invention ensures that the initial state of the system is in the sliding mode, the jitter and the steady-state error of the system are reduced in the whole response process, and meanwhile, a primary item of the tracking error is introduced on the sliding mode surface, so that the linear feedback gain and the convergence speed of the system in the sliding mode are accelerated, and the dynamic performance is improved.

Description

Nonlinear global sliding mode model-free control method based on neural network observer
Technical Field
The invention relates to the technical field of aerospace automation, in particular to a nonlinear global sliding mode model-free control method based on a neural network observer.
Background
The quartz lamp heater has the characteristics of small thermal inertia and low manufacturing cost, and can be combined into heaters with different shapes to heat different positions of the aircraft, so that the quartz lamp heater is widely applied to ground structure thermal tests. However, most of the current researches on the structural thermal tests are carried out through numerical simulation, and the precision of a control algorithm of a quartz lamp heating system is not ideal. Because the model is difficult to establish based on the quartz lamp heating system, the established whole mathematical model contains nonlinear terms and high-order terms, and in addition, the quartz lamp heating system also comprises some external disturbances besides the uncertain terms existing in the system.
The model-free control super-local model has the advantages that the order of the system is reduced, the model is simplified, and uncertainty and external disturbance of the system can be observed through an observer. Therefore, model-free control is applied to a mathematical model of the quartz lamp heater, and a model-free control method meeting the performance requirement of the heater is designed. The iPID control method is a model-free control method, has a simple structure and is easy to debug, but only a proportional term, an integral term and a differential term for parameter adjustment cannot meet the requirements on the control accuracy and the convergence speed of a control system. In addition, sliding mode control is an important branch of modern control theory because it is insensitive to uncertainty and disturbance, considering that a model-free control method incorporating sliding mode acts on the system. The traditional sliding mode control comprises an approaching mode and a sliding mode, the approaching mode can generate a buffeting phenomenon of high-frequency switching, the sliding mode can bring oscillation and overshoot due to discontinuity of a symbol function, the robustness performance of the whole system is directly influenced, and the design of the whole sliding mode surface also needs comprehensive consideration.
Disclosure of Invention
This section is for the purpose of summarizing some aspects of embodiments of the invention and to briefly introduce some preferred embodiments. In this section, as well as in the abstract and the title of the invention of this application, simplifications or omissions may be made to avoid obscuring the purpose of the section, the abstract and the title, and such simplifications or omissions are not intended to limit the scope of the invention.
The present invention has been made in view of the above-mentioned conventional problems.
Therefore, the invention provides a nonlinear global sliding mode model-free control method based on a neural network observer, which can solve the problem that the prior art cannot compensate for the observation error of a hypersonic aircraft pneumatic thermal ground simulation system.
In order to solve the technical problems, the invention provides the following technical scheme: according to the law of conservation of energy, establishing a mathematical model of a hypersonic aircraft aerodynamic heat ground simulation system and a super-local model without model control; constructing a radial cubic b-spline-based adaptive neural network by utilizing a cubic b-spline basis function, and predicting unknown disturbance of the hypersonic aircraft aerodynamic heat ground simulation system; based on high-frequency switching buffeting in a nonlinear global sliding mode surface weakening approaching state; and establishing a nonlinear equivalent control rate and a nonlinear approach rate according to the sliding mode accessibility condition to obtain a nonlinear global sliding mode control rate.
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: the hypersonic aircraft pneumatic heat ground simulation system comprises a non-contact radiation heater, an electric power regulating device and a calorimetric sensor; establishing an input and output energy conservation equation according to the energy conservation law to obtain the current temperature T1And the conduction angle alpha of the triac, as follows,
Figure BDA0002941359900000021
wherein, the left side U of the equationIThe input voltage is the voltage at two ends of the power supply, R is the sum of the resistances of the non-contact radiation heater, alpha is the conduction angle of the bidirectional thyristor, and the right side of the equation is respectively used for internal energy consumed by the non-contact radiation heater and heat energy and heat conduction lost in the convection heat exchange processHeat energy lost in the process, heat energy output by heat radiation effect, c, m, T1、T0A, epsilon and delta t are respectively the specific heat capacity, mass, current temperature, initial temperature, surface area, blackness coefficient and working time of the non-contact radiation heater, and beta, lambda, sigma and F are respectively the convection heat transfer coefficient, heat conduction coefficient, Stefin-Boltzmann constant and angle coefficient.
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: comprises that when the controlled object model is a single-input single-output system, the controlled object model is converted into the model-free control super-local model, as follows,
y(n)＝G+χu(t)
wherein, y(n)The method is expressed as an nth derivative of an output quantity y to time t, n is generally 1 or 2, u is expressed as an input quantity, G is expressed as a set of all unknown disturbances, the unknown disturbances include external disturbances and system internal nonlinear disturbances, and χ is expressed as a non-physical adjustable parameter.
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: according to the model-free control super-local model, dividing two sides of the input and output energy conservation equation by delta t and performing term shift to obtain a mathematical model of the hypersonic aircraft aerodynamic heat ground simulation system, as follows,
Figure BDA0002941359900000031
wherein the content of the first and second substances,
Figure BDA0002941359900000032
is T1The derivative with respect to the time at is,
Figure BDA0002941359900000033
alpha respectively corresponds to y in the model-free controlled super-local model(n)U; while sin2 alpha gives the system a periodicityVibration does not affect the convergence of the entire system, and the term containing sin2 α can be considered as input disturbance, a ε σ FT1 4Can be seen as a higher order output disturbance of the system and thus
Figure BDA0002941359900000034
The sum of all disturbances, which can be seen as both input and output disturbances, corresponds to the G of the hyper-local model, which can be observed by an observer.
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: said cubic b-spline basis functions defining an implied layer in a neural network, comprising,
Figure BDA0002941359900000035
wherein, | | x-oiI is the radial distance, oiIs a cubic b-spline basis function center vector, x is an input vector, hiFor the width of the b-spline basis function,
Figure BDA0002941359900000036
i. j, m and n are all positive integers.
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: further comprising, the radial cubic b-sample strip-based adaptive neural network is as follows,
Figure BDA0002941359900000037
wherein, W*Argmin (g (t)) is g (t) is an ideal weight for the neural network, W*TIs W*The transpose of (a) is performed,
Figure BDA0002941359900000038
is a threshold value;
Figure BDA0002941359900000041
Figure BDA0002941359900000042
wherein the content of the first and second substances,
Figure BDA0002941359900000043
in order to be an observed value of the disturbance,
Figure BDA0002941359900000044
in order to be able to observe the error of the observer,
Figure BDA0002941359900000045
is the weight value under the current observation state,
Figure BDA0002941359900000046
is composed of
Figure BDA0002941359900000047
The transposing of (1).
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: including, the tracking error expression defining the output is as follows,
e(t)＝y*-y
where e is the tracking error, y*Is an output target; obtaining a model-free controller through closed-loop control according to the model-free controlled super-local model, as follows,
Figure BDA0002941359900000048
wherein the content of the first and second substances,
Figure BDA0002941359900000049
is an estimate of the value of G,
Figure BDA00029413599000000410
is y*Is the first order differential of (d), delta (e) is the iPI closed loop feedback control rate, delta (e) ═ Kpe(t)+KiIntegral; in order to attenuate the observed disturbances, an auxiliary controller u is added to the modeless controllerauxThe following, as follows,
Figure BDA00029413599000000411
wherein u isauxIs controlled according to a nonlinear global sliding mode, namely a nonlinear global sliding mode surface.
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: the non-linear global sliding-mode surface comprises,
Figure BDA00029413599000000412
wherein the content of the first and second substances,
Figure BDA00029413599000000413
Figure BDA00029413599000000414
k is the feedback gain, e (0) is the initial error,
Figure BDA00029413599000000415
the first order differential of the slip form surface s is:
Figure BDA00029413599000000416
wherein
Figure BDA0002941359900000051
Is the first differential of e (t).
As a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: obtaining a mathematical relationship between a first order differential of the sliding-mode surface s and the nonlinear global sliding-mode surface, including,
Figure BDA0002941359900000052
according to
Figure BDA0002941359900000053
An equivalent control rate is obtained, as follows,
Figure BDA0002941359900000054
according to sliding mode accessibility
Figure BDA0002941359900000055
The approach rate was obtained as follows,
Figure BDA0002941359900000056
wherein the content of the first and second substances,
Figure BDA0002941359900000057
g0for upper bound of observation error, κ is the adjustable gain,
Figure BDA0002941359900000058
as a preferred scheme of the nonlinear global sliding mode model-free control method based on the neural network observer, the method comprises the following steps: the method also comprises the step of simultaneously fusing the auxiliary controller, the equivalent controller and the approach rate to obtain a nonlinear global sliding mode modeless controller u (t) of the neural network observer of the hypersonic aircraft aerodynamic heat ground simulation system,
Figure BDA0002941359900000059
wherein u isaux＝ueq+ucor。
The invention has the beneficial effects that: aiming at the problems of system uncertainty and external disturbance in a pneumatic thermal ground simulation system of a hypersonic aircraft, the invention introduces a model-control-free super-local model linearization method to carry out reduced-order linearization on the system, secondly, a radial cubic b-sample strip-based adaptive neural network has the property of infinitely approximating any function, designs a radial cubic b-sample strip-based adaptive neural network observer, carries out real-time tracking observation on all unknown disturbance through the radial cubic b-sample strip-based adaptive neural network observer, designs a nonlinear global sliding mode control method to overcome the buffeting phenomenon of a sliding mode surface in the approaching process, ensures that the initial state of the system is in a sliding mode, reduces jitter and steady-state errors in the whole response process of the system, introduces a primary term of tracking errors on the sliding mode surface, accelerates the linear feedback gain and convergence rate of the system on the sliding mode, the dynamic performance is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise. Wherein:
fig. 1 is a schematic working flow diagram of a hypersonic aircraft aerodynamic heat ground simulation system based on a neural network observer nonlinear global sliding mode model-free control method according to an embodiment of the invention;
fig. 2(a) is a schematic diagram of a three-dimensional structure of a hypersonic velocity missile based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention;
fig. 2(b) is a schematic two-dimensional size diagram of a hypersonic velocity missile based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention;
fig. 3(a) is a schematic diagram of a pneumatic thermal finite element simulation of hypersonic missile at 0 ° cruise based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention;
fig. 3(b) is a schematic diagram of a pneumatic thermal finite element simulation of a hypersonic missile at an attack angle of 5 ° cruise based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention;
fig. 3(c) is a schematic diagram of a pneumatic thermal finite element simulation of hypersonic missile at a 10 ° cruise angle of attack based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention;
fig. 4 is a hypersonic missile attack angle 0 ° cruise wall surface average temperature sampling diagram (a) and a data fitting diagram (b), a hypersonic missile attack angle 5 ° cruise wall surface average temperature sampling diagram (c) and a data fitting diagram (d), a hypersonic missile attack angle 10 ° cruise wall surface average temperature sampling diagram (e) and a data fitting diagram (f) based on the nonlinear global sliding mode model-free control method of the neural network observer according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a control principle framework of a nonlinear global sliding-mode model-free control method based on a neural network observer according to an embodiment of the present invention;
fig. 6(a) is an output temperature curve diagram of a hypersonic velocity guided missile attack angle 0 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target (1), of a hypersonic velocity aircraft aerodynamic heating ground simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer (2), a linear global sliding mode model-free control method of a neural network observer (3) and a conventional PID method (4) according to an embodiment of the present invention;
fig. 6(b) is a local enlarged view of a hypersonic velocity guided missile attack angle 0 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target (1), of a hypersonic velocity aircraft aerodynamic heat ground simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer (2), a linear global sliding mode model-free control method of a neural network observer (3), and a conventional PID method (4) according to an embodiment of the present invention;
fig. 6(c) is an output temperature curve diagram of a hypersonic velocity guided missile attack angle 5 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target (1), of a hypersonic velocity aircraft aerodynamic heating ground simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer (2), a linear global sliding mode model-free control method of a neural network observer (3) and a conventional PID method (4) according to an embodiment of the present invention;
fig. 6(d) is a local enlarged view of a hypersonic velocity guided missile attack angle 5 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target (1), of a hypersonic velocity aircraft aerodynamic heat ground simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer (2), a linear global sliding mode model-free control method of a neural network observer (3) and a conventional PID method (4) according to an embodiment of the invention;
fig. 6(e) is an output temperature curve diagram of a hypersonic velocity guided missile attack angle 10 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target (1), of a hypersonic velocity aircraft aerodynamic heat ground simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer (2), a linear global sliding mode model-free control method of a neural network observer (3) and a conventional PID method (4) according to an embodiment of the present invention;
fig. 6(f) is a local enlarged view of a hypersonic velocity guided missile attack angle 10 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target (1), of a hypersonic velocity aircraft aerodynamic heat ground simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer (2), a linear global sliding mode model-free control method of a neural network observer (3) and a conventional PID method (4) according to an embodiment of the invention;
fig. 7(a) is a tracking error curve diagram of a hypersonic velocity guided missile attack angle 0 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention, where the hypersonic velocity vehicle pneumatic thermal ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer, and a conventional PID method (3);
fig. 7(b) is a local enlarged view of a hypersonic velocity guided missile attack angle 0 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target according to an embodiment of the invention, and a hypersonic velocity aircraft pneumatic thermal ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer and a conventional PID method (3);
fig. 7(c) is a tracking error curve diagram of a hypersonic velocity guided missile attack angle 5 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention, where the hypersonic velocity vehicle pneumatic thermal ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer, and a conventional PID method (3);
fig. 7(d) is a local enlarged view of a hypersonic velocity guided missile attack angle 5 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target according to an embodiment of the invention, and a hypersonic velocity aircraft pneumatic thermal ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer and a conventional PID method (3);
fig. 7(e) is a tracking error curve diagram of a hypersonic velocity guided missile attack angle 10 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention, where the hypersonic velocity vehicle pneumatic thermal ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer, and a conventional PID method (3);
fig. 7(f) is a local enlarged view of a hypersonic velocity guided missile attack angle 10 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer, which is taken as a tracking target according to an embodiment of the invention, and a hypersonic velocity aircraft pneumatic thermal ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer and a conventional PID method (3);
fig. 8(b) is an output temperature curve diagram of a hypersonic velocity aircraft aerodynamic surface simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer, which is based on an embodiment of the present invention, and which takes a hypersonic velocity missile attack angle 0 ° cruise wall surface average temperature data fitting graph as a tracking target (1), and under an external disturbance corresponding to the hypersonic velocity missile attack angle 0 ° cruise (fig. 8a), under the nonlinear global sliding mode model-free control method (2) of the neural network observer, the linear global sliding mode model-free control method (3) of the neural network observer, and the conventional PID method (4);
fig. 8(c) is a local enlarged view of a hypersonic velocity aircraft aerodynamic surface simulation system based on a nonlinear global sliding mode model-free control method (2) of a neural network observer, a linear global sliding mode model-free control method (3) of a neural network observer and a traditional PID method (4) under an external disturbance corresponding to 0-degree cruising of the hypersonic velocity missile (fig. 8a) by using a fitting graph of the hypersonic velocity missile attack angle 0-degree cruising wall surface average temperature data as a tracking target (1) based on the nonlinear global sliding mode model-free control method of the neural network observer according to an embodiment of the present invention;
fig. 8(e) is an output temperature curve diagram of a hypersonic velocity aircraft aerodynamic surface simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer, which is based on a fitting graph of the mean temperature data of a hypersonic velocity missile wall surface at an angle of attack of 5 degrees cruise and is taken as a tracking target (1), and under the external disturbance corresponding to the hypersonic velocity missile wall surface at the angle of attack of 5 degrees cruise (fig. 8d), the output temperature curve diagram of the hypersonic velocity aircraft aerodynamic surface simulation system based on the nonlinear global sliding mode model-free control method of the neural network observer (2), the linear global sliding mode model-free control method of the neural network observer (3) and the conventional PID method (4) is taken;
fig. 8(f) is a local enlarged view of a hypersonic aircraft aerodynamic surface simulation system based on a nonlinear global sliding mode model-free control method (2) of a neural network observer, a linear global sliding mode model-free control method (3) of a neural network observer and a traditional PID method (4) under an external disturbance corresponding to a cruise of a hypersonic missile attack angle of 5 ° (fig. 8d) by using a fitting graph of the mean temperature data of a wall surface of a 5 ° cruise hypersonic missile attack angle based on the nonlinear global sliding mode model-free control method of the neural network observer as a tracking target (1) according to an embodiment of the present invention;
fig. 8(h) is an output temperature curve diagram of a hypersonic velocity aircraft aerodynamic surface simulation system based on a nonlinear global sliding mode model-free control method of a neural network observer, which is based on an embodiment of the present invention, and which takes a hypersonic velocity missile attack angle 10 ° cruise wall surface average temperature data fitting graph as a tracking target (1), and under an external disturbance corresponding to a hypersonic velocity missile attack angle 10 ° cruise (fig. 8g), under the nonlinear global sliding mode model-free control method (2) of the neural network observer, the linear global sliding mode model-free control method (3) of the neural network observer, and the conventional PID method (4);
fig. 8(i) is a local enlarged view of a hypersonic aircraft aerodynamic surface simulation system based on a nonlinear global sliding mode model-free control method (2) of a neural network observer, a linear global sliding mode model-free control method (3) of a neural network observer and a traditional PID method (4) under an external disturbance corresponding to a cruise of a hypersonic missile attack angle of 10 ° (fig. 8g) by using a fitting graph of hypersonic missile attack angle 10 ° cruise wall surface average temperature data as a tracking target (1) based on the nonlinear global sliding mode model-free control method of the neural network observer according to an embodiment of the present invention;
fig. 9(a) is a tracking error curve diagram of a hypersonic velocity guided missile attack angle 0 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention, where the hypersonic velocity guided missile attack angle 0 ° cruise corresponds to an external disturbance, and a hypersonic velocity aircraft aerodynamic heat ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer, and a conventional PID method (3);
fig. 9(b) is a local enlarged view of a hypersonic velocity aircraft aerodynamic heat ground simulation system based on a nonlinear global sliding mode model-free control method (1) of a neural network observer, a linear global sliding mode model-free control method (2) of a neural network observer and a traditional PID method (3) under an external disturbance corresponding to 0-degree cruising of the hypersonic velocity missile attack angle, wherein a fitting graph of the hypersonic velocity missile attack angle 0-degree cruising wall surface average temperature data is taken as a tracking target according to the nonlinear global sliding mode model-free control method of the neural network observer;
fig. 9(c) is a tracking error curve diagram of a hypersonic velocity guided missile attack angle 5 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention, where the hypersonic velocity guided missile attack angle 5 ° cruise corresponding external disturbance is used as a tracking target, and a hypersonic velocity vehicle aerodynamic heat ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer, and a conventional PID method (3);
fig. 9(d) is a local enlarged view of a hypersonic aircraft aerodynamic heat ground simulation system based on a nonlinear global sliding mode model-free control method (1) of a neural network observer, a linear global sliding mode model-free control method (2) of a neural network observer and a traditional PID method (3) under an external disturbance corresponding to a cruise of a hypersonic missile at an angle of attack of 5 degrees, wherein a fitting graph of the hypersonic missile angle of attack based on the nonlinear global sliding mode model-free control method of the neural network observer is taken as a tracking target according to the nonlinear global sliding mode model-free control method of the neural network observer in one embodiment of the invention;
fig. 9(e) is a tracking error curve diagram of a hypersonic velocity guided missile attack angle 10 ° cruise wall surface average temperature data fitting graph based on a nonlinear global sliding mode model-free control method of a neural network observer according to an embodiment of the present invention, where the hypersonic velocity guided missile attack angle 10 ° cruise corresponding external disturbance is used as a tracking target, and a hypersonic velocity vehicle aerodynamic heat ground simulation system is based on a nonlinear global sliding mode model-free control method (1) of the neural network observer, a linear global sliding mode model-free control method (2) of the neural network observer, and a conventional PID method (3);
fig. 9(f) is a local enlarged view of a hypersonic aircraft aerodynamic heat ground simulation system based on a nonlinear global sliding mode model-free control method (1) of a neural network observer, a linear global sliding mode model-free control method (2) of a neural network observer and a traditional PID method (3) under an external disturbance corresponding to a cruise of a hypersonic missile attack angle of 10 degrees, wherein a fitting graph of the hypersonic missile attack angle 10-degree cruise wall surface average temperature data is taken as a tracking target according to the nonlinear global sliding mode model-free control method of the neural network observer.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, specific embodiments accompanied with figures are described in detail below, and it is apparent that the described embodiments are a part of the embodiments of the present invention, not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making creative efforts based on the embodiments of the present invention, shall fall within the protection scope of the present invention.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways than those specifically described and will be readily apparent to those of ordinary skill in the art without departing from the spirit of the present invention, and therefore the present invention is not limited to the specific embodiments disclosed below.
Furthermore, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
The present invention will be described in detail with reference to the drawings, wherein the cross-sectional views illustrating the structure of the device are not enlarged partially in general scale for convenience of illustration, and the drawings are only exemplary and should not be construed as limiting the scope of the present invention. In addition, the three-dimensional dimensions of length, width and depth should be included in the actual fabrication.
Meanwhile, in the description of the present invention, it should be noted that the terms "upper, lower, inner and outer" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of describing the present invention and simplifying the description, but do not indicate or imply that the referred device or element must have a specific orientation, be constructed in a specific orientation and operate, and thus, cannot be construed as limiting the present invention. Furthermore, the terms first, second, or third are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
The terms "mounted, connected and connected" in the present invention are to be understood broadly, unless otherwise explicitly specified or limited, for example: can be fixedly connected, detachably connected or integrally connected; they may be mechanically, electrically, or directly connected, or indirectly connected through intervening media, or may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Example 1
Referring to fig. 1 to 5, for a first embodiment of the present invention, a nonlinear global sliding mode model-free control method based on a neural network observer is provided, the method of the present invention is based on a model-free controlled super-local model of a hypersonic aircraft aerodynamic thermal ground simulation system, combines an iPID, a radial cubic b-sample strip-based adaptive neural network, a nonlinear global sliding mode surface, a nonlinear equivalent control rate, and a nonlinear approach rate, and designs a controller u (t) to realize target tracking; referring to fig. 5, a model-free control block diagram of the hypersonic aircraft aerodynamic heat ground simulation system based on the nonlinear global sliding mode of the neural network observer specifically includes:
s1: according to the law of conservation of energy, a mathematical model of a hypersonic aircraft aerodynamic heat ground simulation system and a super-local model without model control are established. It should be noted that the hypersonic aircraft aerodynamic heating ground simulation system includes:
a non-contact radiant heater, an electric power regulating device and a calorimetric sensor;
establishing an input and output energy conservation equation according to the energy conservation law to obtain the current temperature T1And the conduction angle alpha of the triac, as follows,
Figure BDA0002941359900000121
wherein, the left side U of the equationIThe input voltage is the voltage at two ends of the power supply, R is the sum of the resistances of the non-contact radiation heater, alpha is the conduction angle of the bidirectional thyristor, the right side of the equation is respectively used for the internal energy consumed by the non-contact radiation heater, the heat energy lost in the convection heat exchange process, the heat energy lost in the heat conduction process and the heat energy output by the heat radiation effect, and c, m, T1、T0A, epsilon and delta t are respectively the specific heat capacity, mass, current temperature, initial temperature, surface area, blackness coefficient and working time of the non-contact radiation heater, beta, lambda, sigma and F are respectively convection heat transfer coefficient, heat conduction coefficient, Stefin-Boltzmann constant and angle coefficient;
when the controlled object model is a single-input single-output system, the controlled object model is converted into a super-local model without model control, as follows,
y(n)＝G+χu(t)
wherein, y(n)The method is characterized in that the method is expressed as an nth derivative of an output quantity y to time t, n is generally 1 or 2, u is expressed as an input quantity, G is expressed as a set of all unknown disturbances, the set comprises external disturbances and system internal nonlinear disturbances, and χ is expressed as an adjustable parameter with non-physical significance;
according to a super-local model without model control, two sides of an input and output energy conservation equation are divided by delta t and terms are shifted to obtain a mathematical model of the pneumatic thermal ground simulation system of the hypersonic aircraft, as follows,
Figure BDA0002941359900000131
wherein the content of the first and second substances,
Figure BDA0002941359900000132
is T1The derivative with respect to the time at is,
Figure BDA0002941359900000133
alpha corresponds to y in the model-free controlled hyper-local model(n)U; while sin2 alpha brings periodic vibration to the system and does not influence the convergence of the whole system, the term containing sin2 alpha can be regarded as input disturbance, and A epsilon sigma FT1 4Can be seen as a higher order output disturbance of the system and thus
Figure BDA0002941359900000134
The sum of all disturbances, which can be seen as both input and output disturbances, corresponds to the G of the hyper-local model, which can be observed by an observer.
S2: and (3) constructing a radial cubic b-spline-based adaptive neural network by utilizing a cubic b-spline basis function, and predicting unknown disturbance of the hypersonic aircraft aerodynamic heat ground simulation system. It should be noted that, the defining of the cubic b-spline basis function of the hidden layer in the neural network includes:
Figure BDA0002941359900000135
wherein, | | x-oiI is the radial distance, oiIs a cubic b-spline basis function center vector, x is an input vector, hiFor the width of the b-spline basis function,
Figure BDA0002941359900000141
i. j, m and n are positive integers;
the radial cubic b-spline based adaptive neural network is as follows,
Figure BDA0002941359900000142
wherein, W*Argmin (g (t)) is g (t) is an ideal weight for the neural network, W*TIs W*The transpose of (a) is performed,
Figure BDA0002941359900000143
is a threshold value;
Figure BDA0002941359900000144
Figure BDA0002941359900000145
wherein the content of the first and second substances,
Figure BDA0002941359900000146
in order to be an observed value of the disturbance,
Figure BDA0002941359900000147
in order to be able to observe the error of the observer,
Figure BDA0002941359900000148
is the weight value under the current observation state,
Figure BDA0002941359900000149
is composed of
Figure BDA00029413599000001410
Transposing;
the tracking error expression defining the output is as follows,
e(t)＝y*-y
where e is the tracking error, y*Is an output target;
the model-free controller is obtained by closed-loop control based on the model-free controlled super-local model, as follows,
Figure BDA00029413599000001411
wherein the content of the first and second substances,
Figure BDA00029413599000001412
is an estimate of the value of G,
Figure BDA00029413599000001413
is y*Is the first order differential of (d), delta (e) is the iPI closed loop feedback control rate, delta (e) ═ Kpe(t)+Ki∫e(t)dt；
To attenuate the observed disturbances, an auxiliary controller u is added to the modeless controllerauxThe following, as follows,
Figure BDA00029413599000001414
wherein u isauxIs controlled according to a nonlinear global sliding mode, namely a nonlinear global sliding mode surface.
S3: based on the high-frequency switching buffeting in the nonlinear global sliding mode surface weakening approaching state. It should be further noted that the nonlinear global sliding-mode surface includes:
Figure BDA00029413599000001415
wherein the content of the first and second substances,
Figure BDA0002941359900000151
Figure BDA0002941359900000152
k is the feedback gain, e (0) is the initial error,
Figure BDA0002941359900000153
the first order differential of the slip form surface s is:
Figure BDA0002941359900000154
wherein
Figure BDA0002941359900000155
Is the first differential of e (t).
S4: and establishing a nonlinear equivalent control rate and a nonlinear approach rate according to the sliding mode accessibility condition to obtain a nonlinear global sliding mode control rate. It should be further noted that, in this step, obtaining a mathematical relationship between the first differential of the sliding-mode surface s and the nonlinear global sliding-mode surface includes:
Figure BDA0002941359900000156
according to
Figure BDA0002941359900000157
An equivalent control rate is obtained, as follows,
Figure BDA0002941359900000158
according to sliding mode accessibility
Figure BDA0002941359900000159
The approach rate was obtained as follows,
Figure BDA00029413599000001510
wherein the content of the first and second substances,
Figure BDA00029413599000001511
g0for upper bound of observation error, κ is the adjustable gain,
Figure BDA00029413599000001512
simultaneously fusing an auxiliary controller, an equivalent controller and an approach rate to obtain a nonlinear global sliding mode modeless controller u (t) of a neural network observer of a hypersonic aircraft aerodynamic heat ground simulation system,
Figure BDA0002941359900000161
wherein u isaux＝ueq+ucor。
Further, establishing a Lyapunov stability criterion expression, and verifying the convergence of the nonlinear global sliding mode model-free control method based on the neural network observer, wherein the method comprises the following steps:
Figure BDA0002941359900000162
wherein the content of the first and second substances,
Figure BDA0002941359900000163
in order to estimate the error for the weight,
Figure BDA0002941359900000164
adjusting parameter gain for the neural network;
Figure BDA0002941359900000165
Figure BDA0002941359900000171
wherein the content of the first and second substances,
Figure BDA0002941359900000172
satisfy the requirement of
Figure BDA0002941359900000173
Referring to fig. 1, the work flow of the hypersonic aircraft aerodynamic heating ground simulation system mainly comprises the following steps:
(1) collecting aerodynamic thermal data of the hypersonic aircraft: carrying out finite element numerical simulation on the hypersonic missile part through a given flight environment and a given wall material type number; and acquiring the average temperature of the wall surface of the missile at each moment, and performing linear fitting on the sampled data to obtain an expected output value, namely a target value, of the whole hypersonic aircraft pneumatic thermal ground simulation system so as to compare the expected output value with the output value of an actual controller.
(2) The hypersonic aircraft aerodynamic heating ground simulation control system comprises: designing a controller to control a quartz lamp heating system; the target value is loaded into the control board, the conduction angle alpha of the bidirectional thyristor is changed through the control board, the output voltage U is further changed, different output voltage U values correspond to different quartz lamp heating system electric powers P, the actual temperature T1 output by the quartz lamp heater is obtained through the sensor, the tracking error e is obtained through comparison with the target value, the actual temperature T1 is fed back to the controller through a closed loop to adjust the conduction angle alpha of the bidirectional thyristor, and finally tracking control is achieved.
(3) Ground simulation test feedback: and (3) carrying out a heating test on the test piece by using the quartz lamp heater, detecting the performance of the test piece, analyzing the feasibility of the material, selecting the material, and if the material cannot be replaced, carrying out the first step operation again, thereby optimizing the design of the thermal protection system.
Referring to fig. 2, the hypersonic missile is drawn by finite element simulation, and the specific parameters of the missile are as follows: the total length is 7600mm, the projectile body length is 4270mm, the projectile body diameter is 1168.4mm, the included angle of the guidance part is 7 degrees, the radius of the guidance head is 30mm, the included angle is 12.84 degrees, the flying environment is 32km in height, the speed is 6.0 Mach number, and the attack angles are 0 degree, 5 degree and 10 degree cruise.
Referring to fig. 4(b), fitting curves of the data of the average temperature of the 0 ° cruise wall surface of the hypersonic missile are shown as follows,
y*＝1.848*10-5t8-0.001497t7+0.04917t6-0.8402t5+7.977t4-41.34t3+93.82t2+151t+287.4
referring to fig. 4(d), fitting curves of the mean temperature data of the cruising wall surface at the attack angle of the hypersonic missile of 5 degrees are shown as follows,
y*＝4.245*10-6t8-0.0003529t7+0.01254t6-0.2542t5+3.268t4-26.24t3+102.8t2+62.74t+287.5
referring to fig. 4(f), fitting curves of the mean temperature data of the cruise wall surface at the hypersonic missile attack angle of 10 degrees are shown as follows,
y*＝-1.448*10-7t8+1.835*10-5t7-0.0005538t6-0.003846t5+0.4455t4-7.239t3+30.19t2+194.6t+289.1
referring to fig. 5, a block diagram of a nonlinear global sliding mode model-free control based on a neural network observer for a hypersonic aircraft aerodynamic thermal ground simulation system is a further description of the second step hypersonic aircraft aerodynamic thermal ground simulation control system of fig. 1, and the process is as follows: firstly, the error e passes through proportional, integral and differential links of a PID controller; and carrying out real-time tracking observation on all unknown disturbances G by a radial cubic b-spline-based adaptive neural network observer; and secondly, compensating the system observation error through a nonlinear global sliding mode surface, a nonlinear equivalent control rate and a nonlinear approach rate to finally form a controller u (t).
Preferably, the embodiment also needs to be explained in the following, compared with the prior art, the invention discloses a nonlinear global sliding mode model-free control method based on a neural network observer, which aims to realize target tracking by adopting the nonlinear global sliding mode model-free control method based on the neural network observer, design a radial cubic b spline basis adaptive neural network by using a cubic b spline basis function to realize estimation of unknown disturbance of a system, eliminate observation errors by using a nonlinear global sliding mode model-free control algorithm and weaken a high-frequency switching buffeting phenomenon in an approach state, thereby ensuring the dynamic control performance of a hypersonic aircraft aerodynamic heating ground simulation system; the method comprises the following steps that on the basis of a super-local model of a hypersonic aircraft aerodynamic heat ground simulation system without model control, the system is subjected to reduced-order linearization processing; the real-time observation and estimation of system disturbance are realized by combining the iPID and the radial cubic b-sample strip-based adaptive neural network; and the compensation of system observation errors is realized by utilizing a nonlinear global sliding mode surface, a nonlinear equivalent control rate and a nonlinear approach rate, so that the control precision is achieved.
Example 2
Referring to fig. 6 to 9, a second embodiment of the present invention, which is different from the first embodiment, provides a test comparison based on a nonlinear global sliding mode model-free control method of a neural network observer, including:
in the embodiment, a hypersonic velocity missile attack angle 0 degree, 5 degree and 10 degree cruise wall surface average temperature data fitting graph is taken as a tracking target, and a hypersonic velocity aircraft pneumatic thermal ground simulation system is adopted to respectively carry out real-time measurement and comparison on the output temperature and the tracking error of the hypersonic velocity aircraft pneumatic thermal ground simulation system under a neural network observer based nonlinear global sliding mode model-free control method (1), a neural network observer based linear global sliding mode model-free control method (2) and a traditional PID method (3).
And (3) testing environment: the method comprises the steps that a hypersonic aircraft pneumatic thermal ground simulation system is operated on a simulation platform to simulate and track an expected target curve (fig. 4b, 4d and 4f), and the hypersonic aircraft pneumatic thermal ground simulation system is used for testing and obtaining test result data under a neural network observer based nonlinear global sliding mode model-free control method (1), a neural network observer based linear global sliding mode model-free control method (2) and a traditional PID (proportion integration differentiation) method (3); the three methods start the automatic test equipment and use MATLB software to realize the simulation test of the comparison method, the simulation data is obtained according to the test result, each method tests 12 groups of data, each group of data is sampled for 15s, each group of data is calculated to obtain the input temperature and the tracking error of each group of data, and the input temperature and the tracking error are compared with the expected target temperature input by simulation and calculation.
Referring to fig. 8 and 9, wherein the external disturbance is a time-varying resistance R:
R＝3.08×(1+0.0045y*)
cruise output y according to corresponding different angles of attack*。
A linear global sliding mode model-free control method (2) based on a neural network observer is as follows:
Figure BDA0002941359900000191
the specific embodiment has the following parameter settings:
table 1: and a parameter table of a hypersonic aircraft pneumatic thermal ground simulation system.
Figure BDA0002941359900000192
Figure BDA0002941359900000201
Table 2: a nonlinear global sliding mode model-free control method data sheet based on a neural network observer.
Figure BDA0002941359900000202
Table 3: a neural network observer based linear global sliding mode model-free control method parameter table.
Figure BDA0002941359900000203
Table 4: PID control method parameter table.
Figure BDA0002941359900000204
Wherein h is8＝[1 2 5 8 1 3 6 2]，
Figure BDA0002941359900000205
Referring to fig. 6a, 6c and 6e, it can be seen that the three methods can effectively track the wall surface average temperature data fitting curve of the hypersonic missile cruising at different attack angles, the fitting degree of the curve (2) is the best, and when the time is 0-0.2 s, the curve (3) and the curve (4) have a certain difference from a target curve.
Referring to fig. 7, it can be seen that the curve (3) has a large overshoot and has a maximum steady-state error, the curve (2) has a certain steady-state error, and the chattering phenomenon at the later stage is relatively obvious, the curve (1) has a minimum steady-state error, the overshoot at the earlier stage is relatively small, and the chattering phenomenon is smaller than the curve (2).
Referring to fig. 8, it can be seen that under corresponding external disturbance, the three methods can effectively track the wall surface average temperature data fitting curve of the hypersonic missile cruising at different attack angles, the fitting degree of the curve (2) is the best, and when the time is 0-0.2 s, the curve (3) and the curve (4) have a certain difference from a target curve.
Referring to fig. 9, it can be seen that, under the corresponding external disturbance, the curve (3) has a large overshoot, and has a certain steady-state error, and the rapidity is also poor, and the curve (2) has a steady-state error and a large buffeting phenomenon, and the curve (1) has a good rapidity, a minimum steady-state error, and a highest control precision.
Based on the analysis, the control method is superior to other 2 methods in 4 aspects of steady-state error, rapidity, overshoot and control precision, and is beneficial to the nonlinear global sliding mode model-free control method based on the neural network observer, the real-time observation and estimation of system disturbance are realized by combining the iPID and the radial cubic b-sample strip-based adaptive neural network, and the compensation of system observation error is realized by utilizing the nonlinear global sliding mode surface, the nonlinear equivalent control rate and the nonlinear approach rate, so that the control precision is achieved.
It should be noted that the above-mentioned embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.

Claims (10)

1. A nonlinear global sliding mode model-free control method based on a neural network observer is characterized by comprising the following steps: comprises the steps of (a) preparing a mixture of a plurality of raw materials,
according to the law of conservation of energy, establishing a mathematical model of a hypersonic aircraft aerodynamic heat ground simulation system and a super-local model without model control;
constructing a radial cubic b-spline-based adaptive neural network by utilizing a cubic b-spline basis function, and predicting unknown disturbance of the hypersonic aircraft aerodynamic heat ground simulation system;
based on high-frequency switching buffeting in a nonlinear global sliding mode surface weakening approaching state;
and establishing a nonlinear equivalent control rate and a nonlinear approach rate according to the sliding mode accessibility condition to obtain a nonlinear global sliding mode control rate.
2. The neural network observer-based nonlinear global sliding-mode model-free control method according to claim 1, characterized in that: the hypersonic aircraft pneumatic heat ground simulation system comprises a non-contact radiation heater, an electric power regulating device and a calorimetric sensor;
establishing an input and output energy conservation equation according to the energy conservation law to obtain the current temperature T1And the conduction angle alpha of the triac, as follows,
Figure FDA0002941359890000011
wherein, the left side U of the equationIThe input voltage is the voltage at two ends of the power supply, R is the sum of the resistances of the non-contact radiation heater, alpha is the conduction angle of the bidirectional thyristor, the right side of the equation is respectively used for the internal energy consumed by the non-contact radiation heater, the heat energy lost in the convection heat exchange process, the heat energy lost in the heat conduction process and the heat energy output by the heat radiation effect, and c, m, T1、T0A, epsilon and delta t are respectively the specific heat capacity, mass, current temperature, initial temperature, surface area, blackness coefficient and working time of the non-contact radiation heater, and beta, lambda, sigma and F are respectively the convection heat transfer coefficient, heat conduction coefficient, Stefin-Boltzmann constant and angle coefficient.
3. The neural network observer based nonlinear global sliding-mode model-free control method according to claim 1 or 2, characterized in that: comprises the steps of (a) preparing a mixture of a plurality of raw materials,
when the controlled object model is a single-input single-output system, the controlled object model is converted into the model-free control super-local model, as follows,
y(n)＝G+χu(t)
wherein, y(n)The method is expressed as an nth derivative of an output quantity y to time t, n is generally 1 or 2, u is expressed as an input quantity, G is expressed as a set of all unknown disturbances, the unknown disturbances include external disturbances and system internal nonlinear disturbances, and χ is expressed as a non-physical adjustable parameter.
4. The neural network observer-based nonlinear global sliding-mode model-free control method according to claim 3, characterized in that: comprises the steps of (a) preparing a mixture of a plurality of raw materials,
according to the model-free control super-local model, dividing two sides of the input and output energy conservation equation by delta t and performing term shift to obtain a mathematical model of the hypersonic aircraft aerodynamic heat ground simulation system, as follows,
Figure FDA0002941359890000021
wherein the content of the first and second substances,
Figure FDA0002941359890000022
is T1The derivative with respect to the time at is,
Figure FDA0002941359890000023
alpha respectively corresponds to y in the model-free controlled super-local model(n)U; while sin2 alpha brings periodic vibration to the system and does not influence the convergence of the whole system, the term containing sin2 alpha can be regarded as input disturbance, and A epsilon sigma FT1 4Can be seen as a higher order output disturbance of the system and thus
Figure FDA0002941359890000024
The sum of all disturbances, which can be seen as both input and output disturbances, corresponds to the G of the hyper-local model, which can be observed by an observer.
5. The neural network observer-based nonlinear global sliding-mode model-free control method according to claim 4, characterized in that: said cubic b-spline basis functions defining an implied layer in a neural network, comprising,
Figure FDA0002941359890000025
wherein, | | x-oiI is the radial distance, oiIs a cubic b-spline basis function center vector, x is an input vector, hiFor the width of the b-spline basis function,
Figure FDA0002941359890000026
i. j, m and n are all positive integers.
6. The neural network observer-based nonlinear global sliding-mode model-free control method according to claim 5, characterized in that: also comprises the following steps of (1) preparing,
the radial cubic b-spline based adaptive neural network is as follows,
Figure FDA0002941359890000031
wherein, W*Argmin (g (t)) is g (t) is an ideal weight for the neural network, W*TIs W*The transpose of (a) is performed,
Figure FDA0002941359890000032
is a threshold value;
Figure FDA0002941359890000033
Figure FDA0002941359890000034
wherein the content of the first and second substances,
Figure FDA0002941359890000035
in order to be an observed value of the disturbance,
Figure FDA0002941359890000036
in order to be able to observe the error of the observer,
Figure FDA0002941359890000037
is the weight value under the current observation state,
Figure FDA0002941359890000038
is composed of
Figure FDA0002941359890000039
The transposing of (1).
7. The neural network observer-based nonlinear global sliding-mode model-free control method according to claim 6, characterized in that: comprises the steps of (a) preparing a mixture of a plurality of raw materials,
the tracking error expression defining the output is as follows,
e(t)＝y*-y
where e is the tracking error, y*Is an output target;
obtaining a model-free controller through closed-loop control according to the model-free controlled super-local model, as follows,
Figure FDA00029413598900000310
wherein the content of the first and second substances,
Figure FDA00029413598900000311
is an estimate of the value of G,
Figure FDA00029413598900000312
is y*Is the first order differential of (d), delta (e) is the iPI closed loop feedback control rate, delta (e) ═ Kpe(t)+Ki∫e(t)dt；
In order to attenuate the observed disturbances, an auxiliary controller u is added to the modeless controllerauxThe following, as follows,
Figure FDA00029413598900000313
wherein u isauxIs controlled according to a nonlinear global sliding mode, namely a nonlinear global sliding mode surface.
8. The neural network observer-based nonlinear global sliding-mode model-free control method according to claim 7, characterized in that: the non-linear global sliding-mode surface comprises,
Figure FDA00029413598900000314
wherein the content of the first and second substances,
Figure FDA0002941359890000041
0＜γ＜1,η＞0，a＝χ,
Figure FDA0002941359890000042
k is the feedback gain, e (0) is the initial error,
Figure FDA0002941359890000043
the first order differential of the slip form surface s is:
Figure FDA0002941359890000044
wherein
Figure FDA0002941359890000045
Is the first differential of e (t).
9. The neural network observer-based nonlinear global sliding-mode model-free control method according to claim 8, characterized in that: obtaining a mathematical relationship between a first order differential of the sliding-mode surface s and the nonlinear global sliding-mode surface, including,
Figure FDA0002941359890000046
according to
Figure FDA0002941359890000047
An equivalent control rate is obtained, as follows,
Figure FDA0002941359890000048
according to sliding mode accessibility
Figure FDA0002941359890000049
Kappa > 0, the approach rate was obtained as follows,
Figure FDA00029413598900000410
wherein the content of the first and second substances,
Figure FDA00029413598900000411
g0for upper bound of observation error, κ is the adjustable gain,
Figure FDA00029413598900000412
10. the neural network observer-based nonlinear global sliding-mode model-free control method according to claim 9, characterized in that: also comprises the following steps of (1) preparing,
the auxiliary controller, the equivalent controller and the approach rate are fused in a simultaneous mode to obtain a neural network observer nonlinear global sliding mode modeless controller u (t) of the hypersonic aircraft aerodynamic heat ground simulation system,
Figure FDA0002941359890000051
wherein u isaux＝ueq+ucor。
CN202110178100.3A 2021-02-08 2021-02-08 Nonlinear global sliding mode model-free control method based on neural network observer Active CN112965382B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110178100.3A CN112965382B (en) 2021-02-08 2021-02-08 Nonlinear global sliding mode model-free control method based on neural network observer

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110178100.3A CN112965382B (en) 2021-02-08 2021-02-08 Nonlinear global sliding mode model-free control method based on neural network observer

Publications (2)

Publication Number Publication Date
CN112965382A true CN112965382A (en) 2021-06-15
CN112965382B CN112965382B (en) 2021-12-17

Family

ID=76284520

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110178100.3A Active CN112965382B (en) 2021-02-08 2021-02-08 Nonlinear global sliding mode model-free control method based on neural network observer

Country Status (1)

Country Link
CN (1) CN112965382B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113495486A (en) * 2021-08-06 2021-10-12 南京工业大学 Model prediction control method based on extended state observer for structural thermal test
CN114123820A (en) * 2021-11-30 2022-03-01 上海电力大学 Vienna rectifier control method based on BP neural network and fractional order PI
CN114371623A (en) * 2022-01-10 2022-04-19 南京工业大学 Fractional order global sliding mode control method for structural thermal test neural network observer
CN116721134A (en) * 2023-08-10 2023-09-08 武汉能钠智能装备技术股份有限公司四川省成都市分公司 Tracking system and method for non-characteristic target
CN117291230A (en) * 2023-11-23 2023-12-26 湘江实验室 Hammerstein nonlinear system hybrid identification method with closed state

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106406102A (en) * 2016-12-20 2017-02-15 哈尔滨工业大学 Hypersonic aerocraft tracking control method with interference observer
CN109343549A (en) * 2018-10-09 2019-02-15 北京航空航天大学 A kind of Spacecraft Attitude Control, system, medium and equipment
CN110601628A (en) * 2019-08-30 2019-12-20 北方工业大学 Model-free prediction current control method based on extended state observer
CN111459051A (en) * 2020-04-23 2020-07-28 河北工业大学 Discrete terminal sliding mode model-free control method with disturbance observer
CN111865169A (en) * 2020-07-21 2020-10-30 南京航空航天大学 Model-free integral sliding mode control method of ultrasonic motor servo system

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106406102A (en) * 2016-12-20 2017-02-15 哈尔滨工业大学 Hypersonic aerocraft tracking control method with interference observer
CN109343549A (en) * 2018-10-09 2019-02-15 北京航空航天大学 A kind of Spacecraft Attitude Control, system, medium and equipment
CN110601628A (en) * 2019-08-30 2019-12-20 北方工业大学 Model-free prediction current control method based on extended state observer
CN111459051A (en) * 2020-04-23 2020-07-28 河北工业大学 Discrete terminal sliding mode model-free control method with disturbance observer
CN111865169A (en) * 2020-07-21 2020-10-30 南京航空航天大学 Model-free integral sliding mode control method of ultrasonic motor servo system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
PENG GAO 等: "An Adaptive Super Twisting Nonlinear Fractional Order PID Sliding Mode Control of Permanent Magnet Synchronous Motor Speed Regulation System Based on Extended State Observer", 《IEEE ACCESS》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113495486A (en) * 2021-08-06 2021-10-12 南京工业大学 Model prediction control method based on extended state observer for structural thermal test
CN113495486B (en) * 2021-08-06 2023-11-24 南京工业大学 Model prediction control method for structural thermal test based on extended state observer
CN114123820A (en) * 2021-11-30 2022-03-01 上海电力大学 Vienna rectifier control method based on BP neural network and fractional order PI
CN114371623A (en) * 2022-01-10 2022-04-19 南京工业大学 Fractional order global sliding mode control method for structural thermal test neural network observer
CN116721134A (en) * 2023-08-10 2023-09-08 武汉能钠智能装备技术股份有限公司四川省成都市分公司 Tracking system and method for non-characteristic target
CN116721134B (en) * 2023-08-10 2023-10-10 武汉能钠智能装备技术股份有限公司四川省成都市分公司 Tracking system and method for non-characteristic target
CN117291230A (en) * 2023-11-23 2023-12-26 湘江实验室 Hammerstein nonlinear system hybrid identification method with closed state
CN117291230B (en) * 2023-11-23 2024-03-08 湘江实验室 Hammerstein nonlinear system hybrid identification method with closed state

Also Published As

Publication number Publication date
CN112965382B (en) 2021-12-17

Similar Documents

Publication Publication Date Title
CN112965382B (en) Nonlinear global sliding mode model-free control method based on neural network observer
CN112987569B (en) Disturbance upper bound adaptive fractional order global terminal sliding mode model-free control method
CN112987566B (en) Aerodynamic-thermal supercoiled nonlinear fractional order sliding-mode model-free control method
CN112965384B (en) Global nonsingular rapid terminal sliding mode model-free method for adaptive neural network
CN108803336B (en) Design method of self-adaptive LQG/LTR controller of aircraft engine
CN112965364B (en) Modeling and iPID (induced draft identification) method for aerodynamic thermal ground simulation system of hypersonic aircraft
CN113625564B (en) Sliding mode control method based on system model for structural thermal test
CN112859617B (en) iPI model-free adaptive global nonsingular rapid terminal sliding mode control method
CN109883660A (en) A kind of thermal modeling test control method
CN113495486B (en) Model prediction control method for structural thermal test based on extended state observer
Silvestre et al. Aircraft control based on flexible aircraft dynamics
Shen et al. Temperature uniformity control of large-scale vertical quench furnaces for aluminum alloy thermal treatment
CN107944137B (en) Multi-field coupled thermoaeroelastic calculation technology for trajectory state of hypersonic aerocraft
Grigorie et al. How the airfoil shape of a morphing wing is actuated and controlled in a smart way
Lichota et al. Frequency responses identification from multi-axis maneuver with simultaneous multisine inputs
CN115236975B (en) Missile aerodynamic thermal ground finite element analysis and IPD nonlinear sliding mode control method
Zhou Model reduction for nonlinear dynamical systems with parametric uncertainties
CN114706295A (en) Method for controlling fast time-varying thermal load in thermal test for testing strength of aerospace plane
CN113900373B (en) Global nonsingular rapid terminal sliding mode control method for structural thermal test
CN114371623B (en) Fractional order global sliding mode control method for structural thermal test neural network observer
NOLL Aeroservoelasticity
CN113065201B (en) Radiation balance temperature calculation method considering slip correction
Yu et al. Perturbation analysis and control of Mach number 2.4-meter transonic wind tunnel
Rauh et al. Verification and experimental validation of flatness-based control for distributed heating systems
Bai et al. Model Predictive Control of Aerodynamic Heat Simulation Heating System with Discrete-time Linear Extended State Observer

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