Fuel cell air compressor motor full-speed domain depth fuzzy control method and system

Abstract

The invention relates to a fuel cell air compressor motor full-speed domain depth fuzzy control method and system, and belongs to the field of motor control, and the method comprises the steps: defining the input and output variables of a depth fuzzy controller; setting the number of initial depth fuzzy control input and output subsets; fuzzy processing is carried out on input and output variables of the depth fuzzy controller; constructing a fuzzy logic rule under different depth fuzzy control input and output subset combinations; a typical working condition point is selected in consideration of a motor full-speed domain, an interpolation factor is introduced to optimize the number of input and output subsets of a depth fuzzy controller, and a deeper depth fuzzy logic rule is established; and performing defuzzification processing on the output variable of the depth fuzzy controller. The system mainly comprises a depth fuzzy controller module, an SVPWM module, an inverter module and the like, proper model selection is performed on system hardware for reference in order to meet actual control requirements, and the method can effectively improve the motor rotating speed control accuracy and shorten the system response time.

Description

Technical Field

[0001] The present invention relates to the technical field of high-speed permanent magnet synchronous motor control, and in particular to a method and system for full-speed depth fuzzy control of a fuel cell air compressor motor. Background Art

[0002] Permanent magnet synchronous motors have attracted wide attention in applications such as machine tools, air compressors and blowers due to their advantages of small size, light weight, high power density and wide speed range at the same power level. In centrifugal air compressor applications, permanent magnet synchronous motors are often in high-speed operation. Due to changes in air compressor operating conditions and temperature increases, motor parameters change greatly. If the impact of parameter changes on control stability and accuracy is not considered, the motor speed fluctuation will increase, causing increased fluctuations in the air compressor outlet pressure and reduced efficiency.

[0003] Due to the characteristics of strong coupling and nonlinearity, high-speed permanent magnet synchronous motors have high speeds and large temperature rises in practical applications. Parameters such as magnetic flux, resistance, and inductance are prone to changes. The performance of traditional fuzzy control is highly dependent on design experience. It is not robust enough in the face of complex operating conditions of high-speed motors. It is difficult to cope with the nonlinear changes of high-speed permanent magnet synchronous motors under multiple operating conditions such as ultra-slow speed, low speed, high speed, and ultra-high speed, and to accurately control the motor speed. Therefore, it is very important to study a full-speed domain deep fuzzy control method with strong robustness and high precision.

[0004] Traditional fuzzy control simply divides the domain of fuzzy input and output such as motor speed and current into large, relatively large, medium, relatively small, and small ranges, resulting in a large range of fuzzy subsets. When the speed or current fluctuates and changes, it is impossible to fully distinguish the differences between variables. Moreover, due to the small number of membership functions contained, the constructed fuzzy logic rules are limited, and the deeper dynamic characteristics of the motor cannot be extracted. Under complex operating conditions, control responses cannot be made in a timely and accurate manner, resulting in reduced control adaptability and accuracy. Summary of the invention

[0005] In view of this, the object of the present invention is to provide a method and system for deep fuzzy control of a fuel cell air compressor motor in the full speed range, so as to improve the accuracy and reliability of speed control of a high-speed permanent magnet synchronous motor and shorten the system response time.

[0006] To achieve the above object, the present invention adopts the following technical solution: a method for full-speed depth fuzzy control of a fuel cell air compressor motor, comprising the following steps:

[0007] Step S1: define the input and output variables of the deep fuzzy controller;

[0008] Step S2: setting the number of input and output variable subsets of the initial deep fuzzy controller;

[0009] Step S3: fuzzifying the input and output variables of the deep fuzzy controller;

[0010] Step S4: constructing deep fuzzy logic rules under different combinations of deep fuzzy input and output subsets;

[0011] Step S5: In different motor speed domains, an interpolation factor is introduced to optimize the number of input and output subsets of the deep fuzzy controller to establish a deeper level of deep fuzzy logic rules;

[0012] Step S6: Defuzzification processing of the output variables of the deep fuzzy controller.

[0013] In a preferred embodiment: Step 1 specifically includes: the depth fuzzy control method is based on the motor expected speed ω * , observe the speed ω and stator current i d 、i q As input, the reference voltage of the motor stationary coordinate system α and β axis is is the output. Let the input variable be x i , the output variable is y j , x1 is ω * , x2 is ω, x3 is i d 、x4 is i q ; y1 is y2 is

[0014] In a preferred embodiment: the step 2 specifically includes: setting the fuzzy subset range of each input and output variable of the initial deep fuzzy controller according to the number of input and output variables of the deep fuzzy controller:

[0015]

[0016] In the formula, N(x i )、N(y i ) represent the number of each deep fuzzy input and output variable subsets under initial conditions; n and m represent the number of input and output variables of the deep fuzzy controller respectively; They represent the correction parameters of the number of fuzzy input and fuzzy output subsets respectively; the value is affected by the range of the fuzzy domain of each variable, and the value range is generally 0 to 30;

[0017] According to the domain range of each fuzzy variable of the motor, the number of deep fuzzy input variable x1 subsets N(x1) is preliminarily set to 30, the number of x2 subsets N(x2) is 30, the number of x3 and x4 subsets N(x3) and N(x4) are both 10; the number of deep fuzzy output variables y1 and y2 fuzzy subsets N(y1) and N(y2) are both set to 10;

[0018] In a preferred embodiment: Step 3 specifically includes: taking the value range of each variable when the motor is working as its fuzzy domain, and evenly dividing its fuzzy domain according to the number of initial deep fuzzy input and output variable subsets set in step S2; when dividing the deep fuzzy input and output variable domains, it is necessary to ensure that the fuzzy subsets can fully cover the entire working range of the motor from zero to the maximum speed; using a triangular membership function to fuzzify the deep fuzzy input and output variables, and there must be a certain overlap between adjacent membership functions.

[0019] In a preferred embodiment, step 4 specifically includes: when establishing deep fuzzy logic rules, it should be ensured that all possible deep fuzzy input situations correspond to at least one fuzzy rule. According to the initial deep fuzzy input, the number of output variable subsets and the membership function set in steps 2 and 3, the deep fuzzy logic rules are constructed, and the deep fuzzy rule establishment principle is:

[0020] If x1is 1,l and x2isA 2,l …and x i iA i,l then y j iB j,l

[0021] In the formula, A i,l , B j,l Represents the input variable x i and the output variable y j The lth fuzzy subset.

[0022] In a preferred embodiment, step 5 specifically includes: in different operating speed domains of the motor, when the speed error amplitude is greater than the speed error threshold, by introducing the interpolation factor ξ, optimizing the number of deep fuzzy input and output variable subsets in different operating speed domains, establishing deeper deep fuzzy logic rules, and reducing the speed error amplitude. Its different operating speed domains mainly include: starting speed domain (0% to 20% of rated speed), low speed domain (20% to 40% of rated speed), medium and high speed domain (40% to 80% of rated speed), high speed domain (80% to rated speed), etc. The interpolation factor of different speed domains is ξ k , where k is the serial number of the speed domain k = 1, 2, 3, 4; ξ1 is the interpolation factor of the starting speed domain; ξ2 is the interpolation factor of the low speed domain; ξ3 is the interpolation factor of the medium and high speed domain; ξ4 is the interpolation factor of the high speed domain.

[0023] Step S51: In different operating speed domains, the deep fuzzy input variable interpolation factor ξ k (x i ) is expressed as:

[0024]

[0025] Where K is the exponential coefficient; λ>0, δ is a sufficiently small positive number; ε k is the speed error threshold for different operating speed ranges, which is generally set according to the maximum speed within the operating speed range and the motor control accuracy requirements; N(x i ) k,t The number of deep fuzzy input variable subsets before optimization for different operating speed domains; E k is the speed error amplitude in different operating speed ranges.

[0026] The observed speed and expected speed of the motor in different speed ranges are ω k , The motor speed error amplitude in different operating speed ranges is:

[0027]

[0028] Where v is the number of data points; ω k,l , The data of the lth point of the observed speed and expected speed of the motor in different operating speed ranges.

[0029] The number of deep fuzzy input variable subsets after optimization in different operating speed domains is

[0030] Step S52: In different motor speed ranges, the number of deep fuzzy output variable subsets after optimization is N(y j ) k,t+1

[0031]

[0032] In the formula, ρ is generally (0,1); N(y j ) k,t represents the number of initial deep fuzzy output variable subsets in different operating speed domains, and its value is determined by the number of initial deep fuzzy output variable subsets N(y j )Decide.

[0033] Step S53: Based on the number of deep fuzzy input and output variable subsets after optimization in different operating speed domains, the fuzzy domain corresponding to the speed domain is re-divided according to step S3, and the triangular membership function of the fuzzy input and output variables is established. According to the initial deep fuzzy logic rule establishment principle of step S4, it is ensured that the optimized fuzzy input subsets have corresponding deep fuzzy logic rules, and further constructs deeper deep fuzzy logic rules in which the speed error amplitudes in different operating speed domains are all less than the set threshold.

[0034] Step S54: To ensure smooth operation of the transition interval speed, the number of fuzzy input and output subsets in the transition interval is determined by the number of fuzzy input and output subsets after optimization of two adjacent operating speed domains and the range of different speed domains occupied by the transition interval. The number of fuzzy input and output subsets in the transition interval is:

[0035]

[0036] In the formula, N'(x i ) k 、N'(y j ) k Respectively represent the number of fuzzy input and output subsets in the transition interval; μ k , μ k+1 They respectively represent the relative proportions of the transition interval in the adjacent front and rear speed domains.

[0037] Proportional step S55: Based on the number of fuzzy input and output subsets in the transition interval, when dividing the transition interval domain, in order to ensure a smooth change in the speed, the overlapping area between adjacent membership functions of the fuzzy input and output variables in the transition interval should exceed 40%.

[0038] Step S56: To reduce the transition interval speed error threshold, based on the principle of establishing fuzzy logic rules in overlapping areas of optimized fuzzy input and output subsets in adjacent operating speed domains, according to the fuzzy logic rule construction steps, establish transition interval fuzzy logic rules.

[0039] Step S57: Combine the deep fuzzy rules established in step S53 based on the number of deep fuzzy input and output variable subsets optimized in different operating speed domains and the deep fuzzy rules established in the transition interval of step S56, and construct full-speed domain deep fuzzy logic rules in a manner from low to high in the speed operating speed domain.

[0040] In a preferred embodiment, the step 6 specifically includes: according to the different input variables of the deep fuzzy controller, the voltage fuzzy output value U(y) in the stationary coordinate system is inferred by the deeper deep fuzzy logic rules set. j ). The fuzzy output is defuzzified using the centroid method:

[0041]

[0042] Where U * (y j ) is the actual output value of the voltage after defuzzification, U(y j ) is the deep fuzzy output variable y j The fuzzy value of μ(U(y j ) is U(y j)The membership function corresponding to the deep fuzzy output subset.

[0043] The present invention provides a fuel cell air compressor motor full-speed domain deep fuzzy control system, and runs the described fuel cell air compressor motor full-speed domain deep fuzzy control method. The control system includes a deep fuzzy controller module, a coordinate transformation module, an SVPWM module, an inverter module, a current and voltage acquisition module, and a speed and position estimation module.

[0044] In a preferred embodiment: the desired speed ω of the depth fuzzy input * The current sensor detects the three-phase current i abc The current i under the α and β axes of the stationary coordinate system is obtained by sending it to the Clark transformation module. α 、i β ; The current i under the α and β axes in the stationary coordinate system α 、i β and voltage u α 、u β The speed and position estimation module is used to obtain the motor observed speed ω and position information θ; θ and i α 、i β The current i is sent to the Park transformation module to obtain the current i in the rotating coordinate system. d 、i q ; Observe the speed ω and current i d 、i q As the input of the deep fuzzy controller; the reference voltage output of the deep fuzzy controller Connected with the inverse Park transformation module, the voltage under the α and β axes of the stationary coordinate system is obtained Will The signal is sent to the SVPWM module to adjust the duty cycle of the PWM pulse signal and drive the high-speed permanent magnet synchronous motor to operate.

[0045] Compared with the prior art, the present invention has the following beneficial effects: the deep fuzzy controller of the present invention has a fuzzy logic rule network with more input and output membership functions and multi-level nonlinear operation combinations. By utilizing more membership functions, the dynamic characteristics such as speed fluctuation and current amplitude of the motor nonlinear system can be described in more detail. Through the multi-level deep fuzzy logic rule network, the fuzzy input is processed in layers, and deeper dynamic response characteristics under different fuzzy input membership function combinations are extracted, thereby improving the accuracy and reliability of high-speed permanent magnet synchronous motor speed control and shortening the system response time. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A structural block diagram of a permanent magnet synchronous motor system according to a preferred embodiment of the present invention;

[0047] Figure 2 A method flow chart of a preferred embodiment of the present invention;

[0048] Figure 3 is the speed ω of the embodiment of the present invention * Membership function graph;

[0049] Figure 4 is the current i in the embodiment of the present invention q Membership function graph;

[0050] Figure 5 is the output variable of the embodiment of the present invention Membership function graph;

[0051] Figure 6 The curve diagram of the embodiment of the present invention, wherein (a) is the speed change curve diagram, (b) is the current i abc Change curve graph. DETAILED DESCRIPTION

[0052] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0053] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0054] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application; as used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or their combinations.

[0055] like Figure 1 As shown, a fuel cell air compressor motor full-speed domain deep fuzzy control system comprises a deep fuzzy controller module (1), a coordinate transformation module (2, 7, 9), an SVPWM module (3), an inverter module (4), a current and voltage acquisition module (5), and a speed and position estimation (8) module.

[0056] The desired speed ω of the deep fuzzy input * The current sensor detects the three-phase current i abc The current i under the α and β axes of the stationary coordinate system is obtained by sending it to the Clark transformation module. α 、i β ; The current i under the α and β axes in the stationary coordinate systemα 、i β and voltage u α 、u β The speed and position estimation module is used to obtain the motor observed speed ω and position information θ; θ and i α 、i β The current i is sent to the Park transformation module to obtain the current i in the rotating coordinate system. d 、i q ; Observe the speed ω and current i d 、i q As the input of the deep fuzzy controller; the reference voltage output of the deep fuzzy controller Connected with the inverse Park transformation module, the voltage under the α and β axes of the stationary coordinate system is obtained Will The duty cycle of the PWM pulse signal is adjusted by the SVPWM module to drive the high-speed permanent magnet synchronous motor to operate; since the present invention establishes a large number of input and output membership functions and fuzzy logic rules, in order to improve the calculation speed, the controller core board needs to have a main frequency of at least 150MHz, as well as a large memory and storage capacity; in order to improve the control accuracy of the deep fuzzy controller, the signal conditioning board needs to include multiple signal acquisition channels and interfaces such as multi-channel current sampling, voltage sampling, temperature sampling and PWM drive to accurately collect motor current, voltage and other parameters for the deep fuzzy controller to identify; since the selected high-speed motor speed is 125000rpm, its basic frequency is 2kHz, in order to effectively realize the accurate control of the high-speed motor speed, its power device should select a SiC power device with higher efficiency and meeting at least 20kHz.

[0057] like Figure 2 The method flow chart of the preferred embodiment of the present invention includes the following steps:

[0058] Step S1: define the input and output variables of the deep fuzzy controller;

[0059] Step S2: setting the number of input and output variable subsets of the initial deep fuzzy controller;

[0060] Step S3: fuzzifying the input and output variables of the deep fuzzy controller;

[0061] Step S4: constructing deep fuzzy logic rules under different combinations of deep fuzzy input and output subsets;

[0062] Step S5: In different motor speed domains, an interpolation factor is introduced to optimize the number of input and output subsets of the deep fuzzy controller to establish a deeper level of deep fuzzy logic rules;

[0063] Step S6: Defuzzification processing of the output variables of the deep fuzzy controller.

[0064] In step 1: the deep fuzzy control method uses the motor's desired speed ω * , observe the speed ω and stator current i d 、i q As input, the reference voltage of the motor stationary coordinate system α and β axis is is the output. Let the input variable be x i , the output variable is y j , x1 is ω * , x2 is ω, x3 is i d 、x4 is i q ; y1 is y2 is

[0065] In step 2: according to the number of input and output variables of the deep fuzzy controller, set the fuzzy subset range of each input and output variable of the initial deep fuzzy controller:

[0066]

[0067] In the formula, N(x i )、N(y i ) represent the number of each deep fuzzy input and output variable subsets under initial conditions; n and m represent the number of input and output variables of the deep fuzzy controller respectively; They represent the correction parameters of the number of fuzzy input and fuzzy output subsets respectively; the value is affected by the range of the fuzzy domain of each variable, and the value range is generally 0 to 30;

[0068] According to the domain range of each fuzzy variable of the motor, the number of deep fuzzy input variable x1 subsets N(x1) is preliminarily set to 30, the number of x2 subsets N(x2) is 30, the number of x3 and x4 subsets N(x3) and N(x4) are both 10; the number of deep fuzzy output variables y1 and y2 fuzzy subsets N(y1) and N(y2) are both set to 10;

[0069] In step 3: the value range of each variable when the motor is working is taken as its fuzzy domain, and its fuzzy domain is evenly divided according to the number of initial deep fuzzy input and output variable subsets set in step S2. When dividing the domain of deep fuzzy input and output variables, it is necessary to ensure that the fuzzy subset can fully cover the entire working range of the motor from zero to the maximum speed. The deep fuzzy input and output variables are fuzzified using the triangular membership function, and there must be a certain overlap between adjacent membership functions.

[0070] In step 4: when establishing deep fuzzy logic rules, it should be ensured that all possible deep fuzzy input situations correspond to at least one fuzzy rule. Based on the initial deep fuzzy input, the number of output variable subsets and the membership function set in steps 2 and 3, deep fuzzy logic rules are constructed. The principle of establishing deep fuzzy rules is as follows:

[0071] If x1is 1,l and x2isA 2,l …and x i iA i,l then y j iB j,l

[0072] In the formula, A i,l , B j,l Represents the input variable x i and the output variable y j The lth fuzzy subset.

[0073] In step 5: In different operating speed domains of the motor, when the speed error amplitude is greater than the speed error threshold, the interpolation factor ξ is introduced to optimize the number of deep fuzzy input and output variable subsets in different operating speed domains, establish deeper deep fuzzy logic rules, and reduce the speed error amplitude. Its different operating speed domains mainly include: starting speed domain (0% to 20% of rated speed), low speed domain (20% to 40% of rated speed), medium and high speed domain (40% to 80% of rated speed), high speed domain (80% to rated speed), etc. The interpolation factor of different speed domains is ξ k , where k is the serial number of the speed domain k = 1, 2, 3, 4; ξ1 is the interpolation factor of the starting speed domain; ξ2 is the interpolation factor of the low speed domain; ξ3 is the interpolation factor of the medium and high speed domain; ξ4 is the interpolation factor of the high speed domain.

[0074] Step S51: In different operating speed domains, the deep fuzzy input variable interpolation factor ξ k (x i ) is expressed as:

[0075]

[0076] Where K is the exponential coefficient; λ>0, δ is a sufficiently small positive number; ε k is the speed error threshold for different operating speed ranges, which is generally set according to the maximum speed within the operating speed range and the motor control accuracy requirements; N(x i ) k,t The number of deep fuzzy input variable subsets before optimization for different operating speed domains; E k is the speed error amplitude in different operating speed ranges.

[0077] The observed speed and expected speed of the motor in different speed ranges are ωk , The motor speed error amplitude in different operating speed ranges is:

[0078]

[0079] Where v is the number of data points; ω k,l , The data of the lth point of the observed speed and expected speed of the motor in different operating speed ranges.

[0080] The number of deep fuzzy input variable subsets after optimization in different operating speed domains is

[0081] Step S52: In different motor speed ranges, the number of deep fuzzy output variable subsets after optimization is N(y j ) k,t+1

[0082]

[0083] In the formula, ρ is generally (0,1); N(y j ) k,t represents the number of initial deep fuzzy output variable subsets in different operating speed domains, and its value is determined by the number of initial deep fuzzy output variable subsets N(y j )Decide.

[0084] Step S53: Based on the number of deep fuzzy input and output variable subsets after optimization in different operating speed domains, the fuzzy domain corresponding to the speed domain is re-divided according to step S3, and the triangular membership function of the fuzzy input and output variables is established. According to the initial deep fuzzy logic rule establishment principle of step S4, it is ensured that the optimized fuzzy input subsets have corresponding deep fuzzy logic rules, and further constructs deeper deep fuzzy logic rules in which the speed error amplitudes in different operating speed domains are all less than the set threshold.

[0085] Step S54: To ensure smooth operation of the transition interval speed, the number of fuzzy input and output subsets in the transition interval is determined by the number of fuzzy input and output subsets after optimization of two adjacent operating speed domains and the range of different speed domains occupied by the transition interval. The number of fuzzy input and output subsets in the transition interval is:

[0086]

[0087] In the formula, N'(x i ) k 、N'(y j ) k Respectively represent the number of fuzzy input and output subsets in the transition interval; μ k , μ k+1They respectively represent the relative proportions of the transition interval in the adjacent front and rear speed domains.

[0088] Proportional step S55: Based on the number of fuzzy input and output subsets in the transition interval, when dividing the transition interval domain, in order to ensure a smooth change in the speed, the overlapping area between adjacent membership functions of the fuzzy input and output variables in the transition interval should exceed 40%.

[0089] Step S56: To reduce the transition interval speed error threshold, based on the principle of establishing fuzzy logic rules in overlapping areas of optimized fuzzy input and output subsets in adjacent operating speed domains, according to the fuzzy logic rule construction steps, establish transition interval fuzzy logic rules.

[0090] Step S57: Combine the deep fuzzy rules established in step S53 based on the number of deep fuzzy input and output variable subsets optimized in different operating speed domains and the deep fuzzy rules established in the transition interval of step S56, and construct full-speed domain deep fuzzy logic rules in a manner from low to high in the speed operating speed domain.

[0091] In step 6: according to the different input variables of the deep fuzzy controller, the voltage fuzzy output value U(y) in the stationary coordinate system is inferred through the deeper deep fuzzy logic rules set. j ). The fuzzy output is defuzzified using the centroid method:

[0092]

[0093] Where U * (y j ) is the actual output value of the voltage after defuzzification, U(y j ) is the deep fuzzy output variable y j The fuzzy value of μ(U(y j ) is U(y j )The membership function corresponding to the deep fuzzy output subset.

[0094] Taking a high-speed motor with a speed range of 0 to 125,000 rpm as an example, in the field of high-speed motors, firstly, according to the number of initial subsets of the fuzzy input, a triangular membership function is used to reflect the degree of membership. The deep fuzzy control model gradually constructs multi-level fuzzy logic rules by establishing fuzzy logic between the same speed membership interval and different speed errors and the real-time current of the motor within the speed operating range. Under the initial conditions, the simulation results show that the speed error amplitude is about 650 rpm, so the desired speed ω is used. r *Taking the domain division as an example, under the initial condition, the number of speed domain subsets N is 30, the threshold is 130, and λ=0.002 is taken, and the interpolation factor ξ is about 0.62. At the rated speed, the number of fuzzy subsets increases by about 10. When the speed is stable, compare whether the speed error is less than the threshold. If the speed error is greater than the threshold, increase the number of fuzzy subsets by changing the subset range to build a deeper deep fuzzy logic rule network until the speed error drops below the set threshold.

[0095] The speed ω r * The range is 0 to 125000 rpm. The triangular membership function is used to divide the speed area into 6 areas: NB, NM, NS, PS, PM, and PB. According to the above initial fuzzy subset division principle, the speed is divided into 30 layers within this speed range.

[0096] The current i q The range is 0 to 60A. The initial fuzzy domain of current is divided into 10 intervals: ZO, PS1, PS2, PS3, PM1, PM2, PM3, PB1, PB2, and PB3 using triangular membership function;

[0097] The voltage u a The range is 0 to 300 V. The initial fuzzy domain of current is divided into 10 intervals: ZO, PS1, PS2, PS3, PM1, PM2, PM3, PB1, PB2, and PB3 using a triangular membership function;

[0098] In summary, NB means extremely small; NM means small; NS means small; ZO means medium; PS means medium to large; PM means large; PB means large.

[0099] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for deep fuzzy control of a fuel cell air compressor motor at full speed, characterized in that: The following steps are involved: Step S1: define the input and output variables of the deep fuzzy controller; Step S2: setting the number of input and output variable subsets of the initial deep fuzzy controller; Step S3: fuzzifying the input and output variables of the deep fuzzy controller; Step S4: constructing deep fuzzy logic rules under different combinations of deep fuzzy input and output subsets; Step S5: In different motor speed domains, an interpolation factor is introduced to optimize the number of input and output subsets of the deep fuzzy controller to establish a deeper level of deep fuzzy logic rules; Step S6: Defuzzification processing of the output variables of the deep fuzzy controller.

2. A method for full-speed depth fuzzy control of a fuel cell air compressor motor according to claim 1, characterized in that: The step 1 specifically includes: the deep fuzzy control method is based on the motor expected speed ω * , observe the speed ω and stator current i d 、i q As input, the reference voltage of the motor stationary coordinate system α and β axis is is the output; the input variable is x i , the output variable is y j , x1 is ω * , x2 is ω, x3 is i d 、x4 is i q ; y1 is y2 is 3. A method for full-speed depth fuzzy control of a fuel cell air compressor motor according to claim 1, characterized in that: The step 2 specifically includes: setting the fuzzy subset range of each input and output variable of the initial deep fuzzy controller according to the number of input and output variables of the deep fuzzy controller: In the formula, N(x i )、N(y i ) represent the number of each deep fuzzy input and output variable subsets under initial conditions; n and m represent the number of input and output variables of the deep fuzzy controller respectively; They represent the correction parameters of the number of fuzzy input and fuzzy output subsets respectively; the value is affected by the range of the fuzzy domain of each variable, and the value range is generally 0 to 30; According to the domain range of each fuzzy variable of the motor, the number of deep fuzzy input variable x1 subsets N(x1) is preliminarily set to 30, the number of x2 subsets N(x2) is 30, the number of x3 and x4 subsets N(x3) and N(x4) are both 10; the number of fuzzy subsets N(y1) and N(y2) of deep fuzzy output variables y1 and y2 are both set to 10.

4. A method for full-speed depth fuzzy control of a fuel cell air compressor motor according to claim 1, characterized in that: The step 3 specifically includes: taking the value range of each variable when the motor is working as its fuzzy domain, and evenly dividing its fuzzy domain according to the number of initial deep fuzzy input and output variable subsets set in step S2. When dividing the deep fuzzy input and output variable domains, it is necessary to ensure that the fuzzy subsets can fully cover the entire working range of the motor from zero to the maximum speed, and use triangular membership functions to fuzzify the deep fuzzy input and output variables, and there must be a certain overlap between adjacent membership functions.

5. A method for full-speed depth fuzzy control of a fuel cell air compressor motor according to claim 1, characterized in that: The step 4 specifically includes: when establishing the deep fuzzy logic rules, it should be ensured that all possible deep fuzzy input situations correspond to at least one fuzzy rule; based on the initial deep fuzzy input, the number of output variable subsets and the membership function set in steps 2 and 3, the deep fuzzy logic rules are constructed, and the principle of establishing the deep fuzzy rules is: If x1isA 1,l and x2isA 2,l …and x i isA i,l then y j isB j,l In the formula, A i,l , B j,l Represents the input variable x i and the output variable y j The lth fuzzy subset.

6. A method for full-speed depth fuzzy control of a fuel cell air compressor motor according to claim 1, characterized in that: The step 5 specifically includes: in different operating speed domains of the motor, when the speed error amplitude is greater than the speed error threshold, by introducing the interpolation factor ξ, optimizing the number of deep fuzzy input and output variable subsets in different operating speed domains, establishing deeper deep fuzzy logic rules, and reducing the speed error amplitude; the different operating speed domains mainly include: starting speed domain (0% to 20% of the rated speed), low speed domain (20% to 40% of the rated speed), medium and high speed domain (40% to 80% of the rated speed), and high speed domain (80% to the rated speed); the interpolation factor of different speed domains is ξ k , where k is the speed domain number k = 1, 2, 3, 4; ξ1 is the starting speed domain interpolation factor; ξ2 is the low speed domain interpolation factor; ξ3 is the medium and high speed domain interpolation factor; ξ4 is the high speed domain interpolation factor; Step S51: In different operating speed domains, the deep fuzzy input variable interpolation factor ξ k (x i ) is expressed as: Where K is the exponential coefficient; λ>0, δ is a sufficiently small positive number; ε k is the speed error threshold for different operating speed ranges, which is generally set according to the maximum speed within the operating speed range and the motor control accuracy requirements; N(x i ) k,t The number of deep fuzzy input variable subsets before optimization for different operating speed domains; E k is the speed error amplitude in different operating speed ranges; The observed speed and expected speed of the motor in different speed ranges are ω k , The motor speed error amplitude in different operating speed ranges is: Where v is the number of data points; ω k,l , The data of the observed speed and the expected speed of the motor at the lth point in different operating speed ranges; The number of deep fuzzy input variable subsets after optimization in different operating speed domains is Step S52: In different motor speed ranges, the number of deep fuzzy output variable subsets after optimization is N(y j ) k,t+1 In the formula, ρ is generally (0,1); N(y j ) k,t represents the number of initial deep fuzzy output variable subsets in different operating speed domains, and its value is determined by the number of initial deep fuzzy output variable subsets N(y j )Decide; Step S53: Based on the number of deep fuzzy input and output variable subsets after optimization in different operating speed domains, the fuzzy domain corresponding to the speed domain is re-divided according to step S3, and the triangular membership function of the fuzzy input and output variables is established. According to the initial deep fuzzy logic rule establishment principle of step S4, it is ensured that the optimized fuzzy input subsets have corresponding deep fuzzy logic rules, and further a deeper deep fuzzy logic rule is constructed in which the speed error amplitude of different operating speed domains is less than the set threshold value; Step S54: To ensure smooth speed operation in the transition interval, the number of fuzzy input and output subsets in the transition interval is determined by the number of fuzzy input and output subsets after optimization of two adjacent operating speed domains and the range of different speed domains occupied by the transition interval; the number of fuzzy input and output subsets in the transition interval is: In the formula, N'(x i ) k 、N'(y j ) k Respectively represent the number of fuzzy input and output subsets in the transition interval; μ k , μ k+1 They represent the relative proportions of the transition interval in the adjacent front and rear speed domains respectively; Proportional step S55: Based on the number of fuzzy input and output subsets in the transition interval, when dividing the transition interval domain, in order to ensure a smooth change in the speed, the overlapping area between adjacent membership functions of the fuzzy input and output variables in the transition interval should exceed 40%; Step S56: To reduce the transition interval speed error threshold, based on the fuzzy logic rule establishment principle of the optimized fuzzy input and output subsets in the overlapping area of ​​adjacent operating speed domains, the transition interval fuzzy logic rule is established according to the fuzzy logic rule construction step; Step S57: Combine the deep fuzzy rules established in step S53 based on the number of deep fuzzy input and output variable subsets optimized in different operating speed domains and the deep fuzzy rules established in the transition interval of step S56, and construct full-speed domain deep fuzzy logic rules in a manner from low to high in the speed operating speed domain.

7. A method for full-speed depth fuzzy control of a fuel cell air compressor motor according to claim 1, characterized in that: Step 6: According to the different input variables of the deep fuzzy controller, the voltage fuzzy output value U(y) in the stationary coordinate system is inferred through the deeper deep fuzzy logic rules set. j ) ; Defuzzify the fuzzy output using the centroid method: Where U * (y j ) is the actual output value of the voltage after defuzzification, U(y j ) is the deep fuzzy output variable y j The fuzzy value of μ(U(y j )) is U(y j )The membership function corresponding to the deep fuzzy output subset.

8. A fuel cell air compressor motor full speed domain deep fuzzy control system, characterized in that A fuel cell air compressor motor full-speed domain deep fuzzy control method as described in claims 1-7 above is run, and the control system includes a deep fuzzy controller module, a coordinate transformation module, an SVPWM module, an inverter module, a current and voltage acquisition module, and a speed and position estimation module.

9. A fuel cell air compressor motor full speed domain deep fuzzy control system according to claim 8, characterized in that: The desired speed ω of the deep fuzzy input * The current sensor detects the three-phase current i a,b,c The current i under the α and β axes of the stationary coordinate system is obtained by sending it to the Clark transformation module. α 、i β ; The current i under the α and β axes in the stationary coordinate system α 、i β and voltage u α 、u β The speed and position estimation module is used to obtain the motor observed speed ω and position information θ; θ and i α 、i β It is sent to the Park transformation module to obtain the current i in the rotating coordinate system d 、i q ; Observe the speed ω and current i d 、i q As the input of the deep fuzzy controller; the reference voltage output of the deep fuzzy controller Connected with the inverse Park transformation module, the voltage under the α and β axes of the stationary coordinate system is obtained Will The signal is sent to the SVPWM module to adjust the duty cycle of the PWM pulse signal and drive the high-speed permanent magnet synchronous motor to operate.

Images