AERO-ENGINE SURGE ACTIVE CONTROL SYSTEM BASED ON FUZZY CONTROLLER SWITCHING

Abstract

An aero-engine surge active control system based on fuzzy controller switching is provided. The present invention selects a basic controller with the most appropriate current state for switching control according to the operating state of a compressor based on the principle of fuzzy switching, and can realize large-range, adaptive and performance-optimized surge active control. Controllers designed by the present invention realize large-range surge active control through fuzzy switching, so that the effective operating ranges of the controllers are expanded and the reliability of the controllers is improved. The designed controllers can be applied to the active control of surge caused by various causes, so that the adaptability of the controllers is improved and is closer to the actual operating condition of the engine. Some optimization indexes are added in the design process of the controllers, which can realize optimal control under corresponding optimization objectives.

Description

TECHNICAL FIELD

The present invention belongs to the field of aero-engine modeling and control, and relates to an aero-engine surge active control system based on fuzzy controller switching.

BACKGROUND

At present, high-performance aero-engines are developing towards high thrust-weight ratio, high speed and high reliability, which also puts forward higher requirements for aerodynamic stability of compressors. The increase of the thrust-weight ratio of the high-performance engines leads to the increase of the single-stage pressure ratio of the compressors and higher compressor load, so that the aerodynamic stability problem of the aero-engines is gradually prominent and becomes one of the important factors that limit the development of the engines.

In the traditional compressor surge control method, the core idea is to ensure sufficient surge margin at operating points when the operating points of the compressors are designed. In this way, the operating points of the compressors may generate slight disturbance, and can still be maintained in a stable operating space. This method has high reliability, but is often too conservative when the operating points are selected, and sacrifices the engine performance for engine stability. At this time, the superiority of the active surge control method is gradually reflected.

The active control method inhibits the generation and development of unstable states of the compressors such as surge by prestage air injection and adjustment of exhaust volume, so as to expand the stable operating range of the engines and make the operating points of the engines move to a position with high performance. At present, the active surge control methods can be divided into the following three categories: modal based control methods, nonlinear control methods and intelligent control methods. These methods can inhibit the occurrence of the unstable states of the compressors such as surge under certain conditions, but have lower reliability compared with the traditional control methods. If a controller fails, surge may be caused directly, which is one of the reasons that the active control method is not actually applied for a long time. Meanwhile, most of the current active control methods are only targeted at one unstable state or instability cause, and have no good adaptability. In addition, most of the active control methods are often designed only for local operating points, and cannot achieve large-range active surge control. At present, the existing active control methods are based on effective inhibition of the unstable states such as surge, do not consider the performance loss of the engines after the controller takes actions, and are not optimal control strategies under the same performance.

SUMMARY

In view of the problems of low reliability, inadequate adaptability and small operating range in the active control method in the prior art, the present invention provides an aero-engine surge active control system based on fuzzy controller switching.

The present invention adopts the following technical solution:

An aero-engine surge active control system based on fuzzy controller switching is provided. The control system mainly comprises three parts: a basic controller design module, a fuzzy switching module and a control signal fusion module. The design process of each part comprises the following steps:

S1 Designing N_c basic controllers in combination with traditional Lyapunov stability theory according to stability requirements in surge active control, wherein the basic controllers are used for generating a basic control signal u_base of a fuzzy switching controller designed by the present invention, and the above control signal is fused into an actual control signal u_out of the controllers through fuzzy switching in subsequent steps. A specific implementation process is as follows:

S1.1 Designing N_c basic controllers by a Lyapunov stability theory based modal control method, with N_c not less than 2; in each basic controller, using a compressor average flow coefficient Φ and a disturbance first-order mode A as feedback quantities respectively to determine a relationship between the feedback quantities and a control quantity u_c required by a compressor, i.e.:

u _base,1 =k ₁(Φ−Φ₀)

u _base,2 =k ₂ A

u _{base,N c} =k _{N c} (−Φ+Φ₀+A)

in the formulas, k₁, k₂, . . . , k_Nc are controller parameters to be determined; u_base,j represents a basic control signal outputted by an jth basic controller, and Φ₀ is the average flow coefficient of the compressor during steady-state operation;

A determining method of the controller parameter is as follows: conducting linearization based on a traditional compressor Moore-Greitzer model, to obtain the controller parameter in combination with the Lyapunov stability theory.

S1.2 Determining the operating ranges of the N_c basic controllers: when the operating state of the compressor is within the operating ranges of the basic controllers, the basic controllers can ensure the stable operation of the compressor through tip jet, and the operating ranges of the basic controllers can be expressed by the size of the disturbance to the compressor. In addition, the designed basic controllers are required to have different operating ranges. For example, in the design process of the basic controllers, different compressor parameter variables (average flow Φ and first-order mode A) are selected as feedback variables of the basic controllers, and the basic controllers operated in the range of small disturbance and large disturbance are designed respectively.

S1.3 Sequencing the basic controllers according to the size of the disturbance range that can be used for operation, based on the operating ranges of the N_c basic controllers, that is, with the increase of the disturbance to the compressor, the controllers in the operating ranges are converted in this order, wherein the rank of the ith basic controller is recorded as rank_i, and rank_i is an integer from 1 to N_c.

S2 Designing the fuzzy switching module.

The fuzzy switching module obtains a selection trend x of the basic controllers by traditional fuzzy reasoning according to a state variable c_tre of the compressor. The state variable x of the compressor refers to the physical quantity which can reflect the operating state of the compressor. These state variables comprise but are not limited to the compressor average flow and average pressure rise. The selection trend c_tre of the basic controllers is a parameter within a range of 0-1, and is used for representing a weight of a basic controller.

The design of the fuzzy switching module needs to determine a state variable x of module input, fuzzy division of module input and output, fuzzy rules used in fuzzy reasoning, and a defuzzification method; a specific design process is as follows:

S2.1 Determining the state variable x (average flow Φ and the first-order mode A) which can represent the operating state of the compressor, as the input of the fuzzy switching module.

S2.2 Conducting fuzzy division for the state variable x of the input of the fuzzy switching module, and obtaining N_a fuzzy sets in each variable division; conducting fuzzy division for the selection trend c_tre of the outputted basic controller to obtain N_b fuzzy sets; and determining that the input variable belongs to a membership function ƒ_in,(i,C)(x_i) of each fuzzy set and the output variable belongs to a membership function ƒ_out,B(c_tre) of each fuzzy set.

The membership μ_in,(i,C) of the input variable belonging to different fuzzy sets can be calculated in the following mode:

μ_in,(i,C)=ƒ_in,(i,C)(x_i)

wherein x_i is the value of the ith input variable; C is a division fuzzy set; ƒ_in,(i,C) is a membership function that the ith input variable belongs to the fuzzy set C; and μ_in,(i,C) is a membership of the calculated ith input variable in the fuzzy set C.

The membership μ_out,B of the output variable belonging to different fuzzy sets can be calculated in the following mode:

μ_out,B=ƒ_out,B(c_tre)

wherein c_tre is the selection trend of the output; B is a division fuzzy set; ƒ_out,B(c_tre) represents a membership function of output c_tre in an output fuzzy set B; and μ_out,B is a membership of the calculated selection trend c_tre in the fuzzy set B.

The above membership functions ƒ_in,(i,C) and ƒ_out,B generally comprise but are not limited to the following functions:

(1) Gaussion membership function

The Gaussion membership function is determined by two parameters σ and ∈, with the following expression:

$f\left(x_{\mathrm{in}},\sigma ,ϵ\right)=e^{\frac{{\left(x_{\mathrm{in}}-ϵ\right)}^2}{2{\sigma }^2}}$

In the formula, x_in is a variable to be fuzzified, parameter σ is used for adjusting the width of the membership function, parameter E is used for determining the center of a curve, and the calculation result of ƒ is the membership of x_in.

(2) Trapezoidal membership function

The trapezoidal membership function can be determined by four parameters a, b, c, d, with the following expression:

$f\left(x_{\mathrm{in}},a,b,c,d\right)=\{\begin{array}{cc}0,& x\le a\\ \frac{x_{\mathrm{in}}-a}{b-a},& a\le x\le b\\ 1,& b\le x\le c\\ \frac{d-x_{\mathrm{in}}}{d-c},& c\le x\le d\\ 0,& x\ge d\end{array}$

In the formula, x_in is a variable to be fuzzified, parameters a and d are the left and right vertices of the lower bottom of a trapezoid respectively, and parameters b and c are the left and right vertices of the upper bottom of the trapezoid respectively.

(3) Triangular membership function

The triangular membership function is determined by three parameters, with the following expression:

$f\left(x_{in},a,b,c\right)=\{\begin{array}{cc}0,& x\le a\\ \frac{x_{in}-a}{b-a},& a\le x\le b\\ \frac{c-x_{in}}{c-b},& b\le x\le c\\ 0,& x\ge c\end{array}$

In the formula, x_in is a variable to be fuzzified, parameters a and c are the left and right vertices of a base of a triangle, and parameter b represents the upper vertex of the triangle.

S2.3 Establishing a fuzzy rule table of the fuzzy switching module and designing N_rules fuzzy rules.

The fuzzy rules can be explained in if-then format, namely:

If x₁∈C₁,x₂∈C₂, . . . ,x_n∈C_n then c_tre∈B

wherein C₁ represents a fuzzy set to which the first input variable belongs in the fuzzy rule, C represents a fuzzy set to which the second variable belongs, and so on; and B represents a fuzzy set to which output c_tre belongs in the fuzzy rule.

In the fuzzy rules, (x₁∈C₁, x₂ ∈C₂, . . . , x_n∈C_n) is a prior condition of the fuzzy rules, and then a prior membership μ_rule,i of the fuzzy rules can be calculated as:

${\mu }_{\mathrm{rule},i}=\prod _{k=1}^{N_{\mathrm{ins}}}{\mu }_{in,\left(k,C_k\right)}$

wherein μ_in,(k,Ck) is a membership of the input variables calculated in S2.2 in the fuzzy set; N_ins is the number of input variables in the fuzzy rule; and μ_rule,i is a prior membership of the calculated ith fuzzy rule.

S2.4 Conducting defuzzification for the output of the fuzzy switching module so that a result after defuzzification is the calculated controller selection trend c_tre of the fuzzy switching module; conducting defuzzification for output results by a centroid method, with a calculation process as follows:

• • (1) calculating the prior membership μ_rule,i of each fuzzy rule, namely:

${\mu }_{\mathrm{rule},i}=\prod _{k=1}^{N_{\mathrm{ins}}}{\mu }_{in,\left(k,C_k\right)}$

(2) calculating the controller selection trend c_tre by the centroid method:

$c_{\mathrm{tre}}=\frac{{\int }_0^{1}c·\left({\Sigma }_{j=1}^{N_{rules}}{\mu }_{\mathrm{rule},j}·f_{\mathrm{out},B_j}\left(c\right)\right)dc}{{\int }_0^{1}c·\left({\Sigma }_{i=1}^{N_{rules}}{\mu }_{\mathrm{rule},i}·f_{\mathrm{out},B_i}\left(c\right)\right)dc}$

wherein μ_rule,i and μ_rule,j are the calculated prior memberships of ith and jth fuzzy rules respectively; ƒ_out,Bi and ƒ_out,Bj are membership functions of the output fuzzy set B in ith and jth fuzzy rules respectively; c_tre is the calculated controller selection trend.

S3 Designing a control signal fusion module. The input of the control signal fusion module comprises the controller selection trend c_tre and a basic control signal u_base generated by the basic controller, and the output is a control signal u_out after fusion. The control signal fusion module calculates the weight c_tre of each basic controller according to the input controller selection trend w_i, and then conducts weighted fusion for the basic control signal u_base according to the calculated weight, to finally obtain an actual control signal u_out of the controller.

The design of the control signal fusion module needs to determine the controller fusion weight and the weighted fusion method, comprising the following specific steps:

S3.1 Designing the controller fusion weight; and according to the number N_c of the basic controllers, conducting fuzzy division for the selection trend c_tre to obtain a fuzzy set center which can be calculated with the following formula:

$c_i=\frac{1}{N_c-1}\left({\mathrm{rank}}_i-1\right),{\mathrm{rank}}_i=1,2,\dots ,N_c$

In the formula, c_i is the ith fuzzy set center; N_c is the number of the basic controllers; rank_i is a range determined by the ith controller in step S1.3, and the value is an integer from 1 to N_c. For example, when N_c=3, fuzzy set centers are c₁=0, c₂=0.5 and c₃=1.

The weight w_i corresponding to the ith controller can be calculated according to the following formula:

w _i=ƒ_w,i(c_tre)

wherein c_tre is the input of the fusion module, i.e., the controller selection trend; ƒ_w,i is a membership function corresponding to the ith controller, and the membership functions can select the form mentioned in step S2.2; and w_i is the calculated weight corresponding to the ith controller.

S3.2 Conducting weighted fusion for the basic control signal to obtain an actual control signal u_out of the basic controller.

The actual control signal u_out of the basic controller can be obtained by weighted fusion through the following formula:

$u_{out}=\frac{{\Sigma }_{j=1}^{N_c}{w}_j·u_{base,j}}{{\Sigma }_{i=1}^{N_c}{w}_i}$

wherein u_base,j represents the basic control signal outputted by the jth basic controller; w_i and w_i represent the weights corresponding to the ith and jth basic controllers respectively; N_c is the number of the basic controllers; and u_out is the actual output control quantity of the fused controller, i.e., the output control quantity of the fuzzy switching controller designed by the present invention.

The above provides the main design and calculation process of the aero-engine surge active control system based on fuzzy controller switching designed by the present invention.

The present invention has the beneficial effects: the surge active control method based on fuzzy controller switching designed by the patent overcomes the limitations of the surge active control method based on a single controller, solves the problem that the single controller cannot satisfy multiple disturbance and expands the effective operating ranges of the surge active controllers. The method can adjust the weight of each basic controller adaptively according to the size of the disturbance, so that the compressor better adapts to various external disturbances. Thus, an aero-engine axial flow compressor can work stably in a wider operating range, thereby greatly improving the success rate of surge active control and the stability of the compressor and improving the safety and the reliability of the aero-engine.

DESCRIPTION OF DRAWINGS

FIG. 1 is a design flow chart of an aero-engine surge active control system based on fuzzy controller switching;

FIG. 2 is a structural schematic diagram of an aero-engine surge active control system based on fuzzy controller switching;

FIG. 3 is a structural diagram of an aero-engine surge active control system based on fuzzy controller switching in an embodiment of the present invention;

FIG. 4 is a schematic diagram of fuzzy sets and fuzzy rules of input and output of a fuzzy switching module; wherein Fig.(a) is fuzzy set division of an average flow coefficient Φ, Fig.(b) is fuzzy set division of a first-order modal amplitude A, Fig.(c) is fuzzy set division of output selection trend c_tre of a fuzzy switching module, and Fig. (d) shows fuzzy switching rules used by the fuzzy switching module;

FIG. 5 is a schematic diagram of fuzzy set division of a control signal fusion module;

FIG. 6 shows a surge active control process under small disturbance, wherein Fig.(a) shows a change process of a local compressor flow coefficient when a fuzzy switching controller and basic controllers proposed in the present invention implement control in the same operating range; Fig. (b) shows jet flow coefficients generated by the fuzzy switching controller and the basic controllers; Fig.(c) shows a switching signal generated by the fuzzy switching module at this moment;

FIG. 7 shows a surge active control process under moderate disturbance, wherein the meanings of Fig.(a), Fig.(b) and Fig.(c) are the same as those described in FIG. 6; and

FIG. 8 shows a surge active control process under large disturbance, wherein the meanings of Fig.(a), Fig.(b) and Fig.(c) are the same as those described in the previous figure.

DETAILED DESCRIPTION

The present invention is further described below in combination with drawings and embodiments of the present invention.

An aero-engine surge active control system based on fuzzy controller switching is provided. The control system mainly comprises three parts: a basic controller design module, a fuzzy switching module and a control signal fusion module. The design flow chart of the aero-engine surge active control system based on fuzzy controller switching is shown in FIG. 1.

FIG. 2 is a structural schematic diagram of the aero-engine surge active control system based on fuzzy controller switching. It can be seen from the figure that the controller mainly comprises three parts: the fuzzy switching module, the control signal fusion module and basic controllers, wherein the fuzzy switching module further comprises three parts: state variable fuzzification, fuzzy rule reasoning and defuzzification. The fuzzy switching controller generates a controller selection trend c_tre through the fuzzy switching module, and the control signal fusion module obtains an actual controller control signal u_out through weighted fusion according to the selection trend c_tre and a basic control signal u_base.

FIG. 3 is a structural diagram of the aero-engine surge active control system based on fuzzy controller switching in the present embodiment.

A specific implementation process comprises the following specific steps:

S1 Designing basic controllers: designing 3 basic controllers in combination with traditional Lyapunov stability theory according to stability requirements in surge active control, specifically as follows:

S1.1 Designing 3 basic controllers by a Lyapunov stability theory based modal control method, and using a compressor average flow coefficient Φ and a disturbance first-order mode A as feedback quantities respectively to determine a relationship between the feedback quantities and a control quantity u_base,j required by a compressor: designing N_c=3 basic controllers for the basic controllers in FIG. 3 according to design step S1.1; selecting pressure rise ψ_j generated by a compressor jet valve as the control quantity to obtain the control laws of the basic controllers:

u _base,1 =k ₁(Φ−Φ₀)  Controller 1:

u _base,2 =k ₂ A  Controller 2:

u _base,3 =k ₃(−Φ+Φ₀+A)  Controller 3:

wherein Φ is the compressor average flow coefficient; A is a first-order modal amplitude; Φ₀ is the average flow coefficient at a balance point of the compressor; k₁, k₂ and k₃ are controller parameters to be determined. According to the Lyapunov stability theory, it can be determined that in the present embodiment, the values of k₁, k₂ and k₃ are as follows: k₁=−0.1, k₂=0.1 and k₃=0.1.

A determining method of the controller parameter is as follows: conducting linearization based on a traditional compressor Moore-Greitzer model, to obtain the controller parameter in combination with the Lyapunov stability theory. The embodiment of the present invention listed herein takes the Moore-Greitzer model of the compressor as a controlled object. The Moore-Greitzer model of the compressor is shown below:

$\frac{dA}{d\xi }=\frac{3\alpha H}{\left(1+m\alpha \right)W}A\left(1-{\left(\frac{\Phi }{W}-1\right)}^2-\frac{1}{4W^2}A\right)$ $\frac{d\Phi }{d\xi }=\frac{H}{l_c}\left(-\frac{\Psi -{\Psi }_{C0}}{H}+\frac{3}{2}\left(\frac{\Phi }{W}-1\right)\left(1-\frac{1}{2W^2}A\right)-\frac{1}{2}{\left(\frac{\Phi }{W}-1\right)}^3+1\right)$ $\frac{d\Psi }{d\xi }=\frac{1}{4B^2{l}_c}\left(\Phi \left(\xi \right)-{\Phi }_T\left(\xi \right)\right)$

In the equations, A(ξ) is a first-order modal amplitude, Φ(ξ) is the compressor average flow coefficient, ψ(ξ) is the average pressure rise coefficient of the compressor and Φ_T(ξ) is the average flow coefficient of a downstream valve of the compressor; in the equations, other parameters are inherent parameters of the compressor, and select the following values here:

ψ_C0=0.30,H=0.14,W=0.25,l_C=8.0,α=1/3.5 and m=1.75.

S1.2 Determining the operating ranges of the 3 basic controllers: when the operating state of the compressor is within the operating ranges of the basic controllers, the basic controllers can ensure the stable operation of the compressor through tip jet, and the operating ranges of the basic controllers can be expressed by the size of the disturbance to the compressor. The performance characteristics of the above three basic controllers are shown in Table 2.

TABLE 2
Performance Characteristics of Basic Controllers
Controllers               Performance Characteristics
Controller 1 (average     Applicable to small disturbance conditions
flow coefficient Φ)       with small flow coefficient change and slow
                          development, with less jet quantity
Controller 2 (first-order Applicable to large disturbance state, with
modal amplitude A)        larger jet quantity
Controller 3              Applicable to moderate disturbance state,
(comprehensive            with less and smooth fluctuation of controlled
feedback (Φ, A))          quantity and moderate jet quantity

S1.3 Sequencing the basic controllers according to the size of the disturbance range that can be used for operation, based on the operating ranges of the 3 basic controllers, that is, with the increase of the disturbance to the compressor, the controllers in the operating ranges are converted in this order, wherein the rank of the ith basic controller is recorded as rank, and rank_i is an integer from 1 to 3. It can be seen from Table 2 that the 3 basic controllers have respective operating ranges. With the increase of the disturbance to the compressor, the controllers in the operating ranges are converted from the controller 1 to the controller 3, and then converted to the controller 2. Therefore, the range of the basic controllers can be recorded as follows:

rank₁=1  Controller 1:

rank₂=3  Controller 2:

rank₃=2  Controller 3:

S2 Designing the fuzzy switching module.

FIG. 4 is a schematic diagram of fuzzy sets and fuzzy rules of input and output of the fuzzy switching module, which reflects the design process of the fuzzy switching module, wherein Fig.(a) is fuzzy set division of an average flow coefficient Φ, Fig.(b) is fuzzy set division of a first-order modal amplitude A, Fig.(c) is fuzzy set division of output controller selection trend c_tre of the fuzzy switching module, and Fig.(d) shows fuzzy switching rules used by the fuzzy switching module.

The fuzzy switching module obtains a selection trend x of the basic controllers by traditional fuzzy reasoning according to a state variable c_tre of the compressor. The state variable x of the compressor comprises but is not limited to the compressor average flow c and average pressure rise ψ. The selection trend c_tre of the basic controllers is a parameter within a range of 0-1, and is used for representing a weight of a basic controller.

S2.1 Determining the average flow Φ and the first-order mode A which can represent the operating state of the compressor, as the input of the fuzzy switching module.

S2.2 Conducting fuzzy division for the state variable x inputted by the fuzzy switching module, as shown in FIG. 4(a):

Fuzzy division is conducted for the average flow coefficient Φ and the first-order modal amplitude A to obtain N_a=5 fuzzy sets. A reminder membership function and a triangular membership function are used in membership functions. The parameters of the selected membership functions are shown in Table 3 and Table 4 respectively (in order to unify the format, the triangular membership function is regarded as a trapezoidal membership function with two endpoints on upper bottom overlapping). Fuzzy set division of the average flow coefficient Φ and the first-order modal amplitude A is shown in Fig. (a) and Fig. (b) respectively.

TABLE 3
Membership Function Parameters of Fuzzy
Sets of Average Flow Coefficients
	Fuzzy sets	a	b	c	d
	Z	0	0	0	0.15
	PZ	0.1	0.25	0.25	0.4
	PS	0.25	0.4	0.4	0.43
	PM	0.43	0.6	0.6	0.63
	PB	0.63	0.8	1	1

TABLE 4
Membership Function Parameters of Fuzzy
Sets of First-Order Modal Amplitude
	Fuzzy sets	a	b	c	d
	Z	0	0	0	0.15
	PZ	0.05	0.18	0.18	0.34
	PS	0.2	0.35	0.35	0.55
	PM	0.4	0.55	0.55	0.75
	PB	0.63	0.75	1	1

Fuzzy division is conducted for the selection trend c_tre of the output of the fuzzy switching module to obtain N_b=5 fuzzy sets. The triangular membership function is used in the membership function. The parameters of the selected membership function are shown in Table 5. Fuzzy set division of the selection trend c_tre is shown in Fig.(c).

TABLE 5
Membership Function Parameters of Fuzzy Sets of Selection trend
	Fuzzy sets	a	b	c
	S	0	0	0.25
	MS	0	0.25	0.5
	M	0.25	0.5	0.75
	MB	0.5	0.75	1
	B	0.75	1	1

S2.3 Establishing a fuzzy rule table, as shown in FIG. 4(d):

According to the performance characteristics of the basic controllers, it can be seen that as the disturbance to the compressor is continuously increased, the controllers are gradually transitioned from average flow coefficient feedback to comprehensive feedback, and then gradually transitioned to first-order modal amplitude feedback according to the operating ranges of the basic controllers. Meanwhile, with the gradual increase of the selection trend c_tre, the used basic controllers are gradually transitioned from the average flow coefficient feedback to the comprehensive feedback, and then gradually transitioned to the first-order modal amplitude feedback. Thus, the design of the fuzzy rules can follow the principle that the greater the compressor disturbance is, the greater the selection trend c_tre is, that is, with the continuous increase of the average flow coefficient Φ and the first-order modal amplitude A, the selection trend c_tre is gradually transitioned from the fuzzy set S to the fuzzy set B. The fuzzy rule table corresponding to the above principle is shown in FIG. 4(d).

S2.4 Conducting defuzzification for the output of the fuzzy switching module to obtain the selection trend c_tre. The process of defuzzification is illustrated by a specific example here.

The average flow coefficient Φ=0.275 and the first-order mode A=0.387 are taken as an example:

(1) Calculating the prior membership μ_rule,i of each fuzzy rule

The following fuzzy rule is taken as an example, i.e.

if Φ∈PZ,A∈PS then c_tre∈MS

The rule of the prior membership can be calculated as

$\begin{array}{c}{\mu }_{\mathrm{rule},1}=\prod _{k=1}^2{\mu }_{in,\left(k,C_k\right)}={\mu }_{in,\left(1PZ\right)}\times {\mu }_{in,\left(2PS\right)}\\ =f_{in,\left(1PZ\right)}\left(\Phi \right)\times f_{in,\left(2PS\right)}\left(A\right)=0.833\times 0.815\\ =0.679\end{array}$

(2) Calculating the controller selection trend c_tre by the centroid method

After the prior membership ρ_rule,i of each fuzzy rule is determined, the controller selection trend c_tre can be calculated according to the defuzzification method by the centroid method in S2.4.

$c_{\mathrm{tre}}=\frac{{\int }_0^{1}c·\left({\Sigma }_{j=1}^{25}{\mu }_{\mathrm{rule},j}·f_{\mathrm{out},B_j}\left(c\right)\right)dc}{{\int }_0^{1}c·\left({\Sigma }_{i=1}^{25}{\mu }_{\mathrm{rule},i}·f_{\mathrm{out},B_i}\left(c\right)\right)dc}=0.286$

S3 Designing a control signal fusion module. FIG. 5 is a schematic diagram of fuzzy set division of the control signal fusion module.

S3.1 Designing a controller fusion weight. In FIG. 5, the fuzzy set P represents the weight of the average flow coefficient feedback, the fuzzy set PA represents the weight of the comprehensive feedback, and the fuzzy set A represents the weight of the first-order modal amplitude feedback, all of which use the triangular membership functions. In the present embodiment, three basic controllers are used, i.e., N_c=3. According to step S3.1, fuzzy set centers can be determined as c₁=0, c₂=0.5 and c₃=1 respectively. Membership function parameters of the fuzzy sets of the control signal fusion module are shown in Table 6.

TABLE 6
Membership Function Parameters of Fuzzy
Sets of Control Signal Fusion Module
	Fuzzy sets	a	b	c
	P	0	0	0.3
	PA	0.2	0.5	0.8
	A	0.7	1	1

S3.2 Conducting weighted fusion for the basic control signal, and calculating weights corresponding to each basic controller according to the fuzzy membership of the control signal fusion module labeled with 6 in S3.1 to obtain the actual output control quantity of the controller after fusion, i.e., the output control quantity u_out of the fuzzy switching controller designed by the present invention. Calculation results are shown in FIG. 6, FIG. 7 and FIG. 8.

The process of weighted fusion of the control signal is further illustrated by an example here:

The selection trend c_tre=0.286 of the controller is calculated in step S2.4 of the detailed description. According to the method in S3.2, the weights corresponding to the basic controllers can be obtained as follows:

w ₁=ƒ_w,1(c_tre)=0.0457

w ₂=ƒ_w,2(c_tre)=0

w ₃=ƒ_w,3(c_tre)=0.288

Then, the outputs of the basic controllers are respectively

u _base,1 =k ₁(Φ−Φ₀)=0.737

u _base,2 =k ₂ A=1.285

u _base,3 =k ₃(−Φ+Φ₀+A)=1.369

The controller output after fusion is

$u_{out}=\frac{{\sum }_{j=1}^3{w}_j·u_{base,j}}{{\sum }_{i=1}^3{w}_i}=\frac{0.0457\times 0.737+0\times 1.285+0.288\times 1.369}{0.0457+0+0.288}=1.283$

The simulation calculation results of this implementation case are shown in FIG. 6, FIG. 7 and FIG. 8: FIG. 6 shows a surge active control process under small disturbance. This disturbance is generated by smaller initial disturbance, wherein Fig. (a) shows a change process of a compressor flow coefficient when a fuzzy switching controller and basic controllers proposed in the present invention implement control; Fig. (b) shows jet flow coefficients generated by the fuzzy switching controller and the basic controllers; Fig.(c) shows a selection trend c_tre generated by the fuzzy switching module at this moment. It can be seen from the figure that under the action of the fuzzy switching controller, the control effect is consistent or even slightly better than that of the basic controllers, and the jet quantity of the jet valve can also be basically consistent. Meanwhile, the selection trend c_tre generated by the fuzzy switching module is basically kept at a low level, that is, the controller which uses the average flow coefficient Φ as the feedback quantity. This is consistent with the performance characteristics and the operating ranges of the basic controllers in the design process.

FIG. 7 shows a surge active control process under moderate disturbance. This disturbance is generated by larger initial disturbance, wherein the meanings of Fig.(a), Fig.(b) and Fig.(c) are the same as those described in FIG. 6. The fuzzy switching controller can achieve the surge active control effect consistent with the basic controllers under moderate disturbance. However, the jet quantity generated by the jet valve used by the fuzzy switching controller is significantly less than that of the basic controllers, which can reflect that the fuzzy switching controller can achieve a more optimized or even optimal control strategy under the required performance indexes. The selection trend c_tre of the controller indicates that at this moment, the controller is gradually transitioned from comprehensive feedback to average flow coefficient feedback, which also conforms to the performance characteristics and the operating ranges of the basic controllers in the design process.

FIG. 8 shows a surge active control process under large disturbance. This disturbance is generated by placing distortion plates in front of the compressor, wherein the meanings of Fig.(a), Fig.(b) and Fig.(c) are the same as those described in the previous figure. The distortion plates can significantly change the inlet flow field of the compressor, which has great influence on the stable operation of the compressor. Under this condition, the fuzzy switching controller can achieve the same or even better performance than the basic controllers in the process of surge active control. Meanwhile, the jet quantity generated by the jet valve is also significantly reduced after comparison, so that the control process has less influence on the operating performance of the engine. In this state, the change process of the controller selection trend c_tre among the basic controllers can be clearly observed. The selection trend is decreased gradually with the size of the disturbance to the compressor, which indicates that the controller is gradually transitioned from the first-order modal amplitude feedback to the comprehensive feedback and then to the average flow coefficient feedback at this moment. It can also be seen from the above process that the surge active control method based on fuzzy switching can control the stable operation of the compressor under the conditions of large disturbance and multiple surge causes, and the reliability and the adaptability of the controllers are significantly improved.

The above embodiments only express the implementation of the present invention, and shall not be interpreted as a limitation to the scope of the patent for the present invention. It should be noted that, for those skilled in the art, several variations and improvements can also be made without departing from the concept of the present invention, all of which belong to the protection scope of the present invention.

Claims

1. An aero-engine surge active control system based on fuzzy controller switching, the control system mainly comprising three parts: a basic controller design module, a fuzzy switching module and a control signal fusion module, wherein the design process of each part comprises the following steps: S1 designing Nc basic controllers according to stability requirements in surge active control, wherein the basic controllers are used for generating a basic control signal ubase of a fuzzy switching controller designed by the present invention, and an implementation process is as follows:S1.1 designing Nc basic controllers by a Lyapunov stability theory based modal control method, with Nc not less than 2; in each basic controller, using a compressor average flow coefficient Φ and a disturbance first-order mode A as feedback quantities respectively to determine a relationship between the feedback quantities and a control quantity uc required by a compressor, i.e.: ubase,1=k1(Φ−Φ0)ubase,2=k2A ubase,Nc=kNc(−Φ+Φ0+A)

2. The aero-engine surge active control system based on fuzzy controller switching according to claim 1, wherein in step S1.1, a determining method of the controller parameter is as follows: conducting linearization based on a traditional compressor Moore-Greitzer model, to obtain the controller parameter in combination with the Lyapunov stability theory.

3. The aero-engine surge active control system based on fuzzy controller switching according to claim 1, wherein in step S2, the compressor state variable x comprises average flow and average pressure rise of the compressor.

4. The aero-engine surge active control system based on fuzzy controller switching according to claim 1, wherein in step S2, the selection trend ctre of the basic controller is within a range of 0-1.