Water-air dual-purpose non-inductive motor direct thrust control method based on hybrid observation strategy

Abstract

The invention provides a water-air dual-purpose non-inductive motor direct thrust control method based on a hybrid observation strategy. The water-air dual-purpose non-inductive motor direct thrust control method comprises the following steps: estimating an electrical angle and an electrical angular speed of a motor rotor by adopting a speed-division non-inductive hybrid observation strategy; based on a high and low speed observation algorithm smooth switching mechanism of dynamic hysteresis, smooth fusion of two algorithm outputs in a dynamic hysteresis interval is realized through a weighted fusion algorithm; direct thrust control models are respectively constructed for water-air cross-medium scenes, and direct thrust is controlled by using a fuzzy-adaptive fractional order sliding mode controller. The hybrid observation strategy gives consideration to high-speed and low-speed working conditions, and the observation precision is improved by more than 30%; through a smooth switching mechanism based on dynamic hysteresis, high and low speed switching jitter is eliminated, and angle and speed estimation fluctuation is reduced by more than 40%; the accuracy of flight control response is improved and the control precision is optimized through a non-inductive motor direct thrust control method in a water-air different-domain scene; the thrust response speed is increased by more than 25%, and the control error is reduced to be within 5%.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, and in particular to a direct thrust control method for a dual-purpose (water and air) sensorless motor based on a hybrid observation strategy. Background Technology

[0002] With the rapid development of emerging technologies such as amphibious unmanned aerial vehicle (UAV) platforms, sensorless motors are gradually becoming core power components in such applications due to their simple structure, low maintenance costs, and high reliability. Especially in amphibious scenarios, sensorless motors need to operate stably under complex conditions of low speed and high torque underwater and high speed and low torque in the air, placing higher demands on the robustness and adaptability of their control strategies.

[0003] To achieve seamless operation, existing technologies typically rely on a single observation algorithm to obtain the motor's electrical angle and speed, thereby controlling the motor. However, this approach struggles to meet the performance requirements of both high-speed and low-speed scenarios in practical applications, and it also fails to adapt to the complex operating conditions of cross-domain UAVs (unmanned aerial vehicles) operating both in water and air. Specifically, existing observation algorithms designed for high-speed motors suffer from insufficient observation accuracy in low-speed scenarios due to signal noise and weak back electromotive force; while optimized observation algorithms designed for low-speed motors have high computational complexity, requiring significant hardware computing power and limiting real-time performance in high-speed scenarios. Because the motor needs to operate and be controlled in a cross-medium environment, the limitations of a single observation algorithm are particularly pronounced, making it difficult to meet the performance requirements under complex operating conditions.

[0004] Secondly, regarding control methods, traditional sensorless motors often employ speed control strategies. This involves the flight control system (FCS) calculating the target attitude and then mapping the required thrust to the motor's PWM signal or speed command for output. However, this mapping process inevitably introduces errors, especially in amphibious applications with frequent dynamic responses or significant load variations. The thrust-to-speed mapping error accumulates over time, causing the actual output thrust to gradually deviate from the flight control command, thus affecting the system's control accuracy and stability. Furthermore, traditional speed control methods cannot directly respond to the flight control's thrust commands, increasing the complexity of the control chain and hindering the achievement of high-precision attitude control and rapid dynamic response.

[0005] In summary, the application of existing sensorless motor technology in amphibious scenarios faces challenges such as insufficient robustness of observation algorithms and speed control mapping errors. These limitations not only restrict the motor's performance under complex operating conditions but also constrain its efficient and stable operation on amphibious platforms. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a direct thrust control method for a dual-purpose (water and air) sensorless motor based on a hybrid observation strategy. This invention can balance robustness and computational efficiency under high and low speed conditions, and realize direct thrust control of the motor technology.

[0007] The technical solution of this invention is: a direct thrust control method for a dual-purpose (water and air) sensorless motor based on a hybrid observation strategy, comprising the following steps:

[0008] S1) A rotor position and speed observation strategy using speed-segment sensorless hybrid control is adopted, and the low-speed observation algorithm and the medium- and high-speed observation algorithm are dynamically selected according to the motor mechanical speed to estimate the motor rotor electrical angle and electrical angular velocity.

[0009] S2) Based on the smooth switching mechanism of dynamic hysteresis, the high-speed and low-speed observation algorithm is switched, and the output of the two algorithms within the dynamic hysteresis interval is smoothly fused through the weighted fusion algorithm.

[0010] S3) Direct thrust control models are constructed for water-air cross-medium scenarios. In the air, a thrust formula based on Bernoulli's equation, momentum theorem and leaf element theory is used. In the water, a correction factor for the influence of the underwater environment is introduced to adjust the direct thrust in the air to obtain the direct thrust of the motor in the water.

[0011] S4) Direct thrust is controlled using a fuzzy-adaptive fractional sliding membrane controller.

[0012] Preferably, in step S1), when the motor is in the air at medium to high speed, a medium to high speed observation algorithm combining the Luenberger state observation algorithm and the phase-locked loop (PLL) is used; when the motor is in the water at low speed, a low speed observation algorithm combining the super torsional sliding mode algorithm and the Kalman filter algorithm is used.

[0013] Preferably, in step S1), a low-speed observation algorithm combining the super torsional sliding mode algorithm and the Kalman filter algorithm is used to estimate the electric angle and electric angular velocity of the motor rotor, specifically as follows:

[0014] Construct the motor voltage equation in the α-β coordinate system and define the sliding surface S. α S β For the current error, a sliding mode control term is generated using a super torsional control law, where the control law satisfies:

[0015]

[0016] In the formula, k1 and k2 are control gains; v α v β These are auxiliary state variables; These represent the rates of change of the auxiliary state variables; sign represents the sign function.

[0017] The extended Kalman filter algorithm uses the electric angle and electric angular velocity of the motor as state variables. It optimizes the back EMF observations through state equations and measurement equations to achieve accurate estimation of rotor position and speed.

[0018] Preferably, in step S2), a dynamic hysteresis interval is introduced, and the hysteresis boundary is dynamically adjusted according to the trend of speed increase or decrease. Within the hysteresis interval, the outputs of the two algorithms are fused using a nonlinear weighted fusion algorithm based on exponential decay.

[0019] Preferably, in step S2), the outputs of the two algorithms are fused using a nonlinear weighted fusion algorithm based on exponential decay, specifically as follows:

[0020]

[0021] In the formula, To estimate the electrical angle; To estimate the electric angular velocity; For fusion factor; The hysteresis loop midpoint is represented by δ, and the attenuation coefficient is ω′. low ,ω′ high The adjusted hysteresis boundary; ω m (k) represents the motor speed at the k-th sampling time; These correspond to the estimated electrical angle and estimated electrical angular velocity obtained by the super torsional sliding mode algorithm and the Kalman filter algorithm, respectively. The estimated electrical angle and estimated electrical angular velocity are obtained by the Luenberger observation algorithm and the phase-locked loop (PLL) algorithm, respectively.

[0022] Preferably, in step S3), the direct air thrust T of the motor... air for:

[0023]

[0024] In the formula, T air ψ is the direct thrust of the motor in the air; k is the in-air propulsion efficiency; p is the number of pole pairs of the motor; ψ is the flux linkage of the permanent magnet; I q ω is the q-axis current; ω is the angular velocity of the propeller. C is the average chord length of the propeller blades; L This is the lift coefficient.

[0025] Preferably, in step S3), the direct thrust T of the motor in the water... water for:

[0026]

[0027] In the formula, Kenv κ is the correction factor. ′ For underwater propulsion efficiency.

[0028] Preferably, in step S4), the fuzzy-adaptive fractional-order sliding mode controller includes a fuzzy inference module, a fractional-order sliding mode controller, and a zero-speed current suppression module ZSCR; the fuzzy inference module dynamically adjusts the sliding mode controller parameters according to the current system mode and error amplitude; the fractional-order sliding mode controller constructs the sliding surface s(t) and control law u(t) according to the output parameters λ, η, β, α of the fuzzy inference module; and the zero-speed current suppression module ZSCR implements a limiting strategy on the output control quantity i(t).

[0029] Preferably, in step S4), the fractional-order sliding mode controller constructs the sliding surface s(t) and the control law u(t) based on the output parameters λ, η, β, and α of the fuzzy inference module, specifically as follows:

[0030]

[0031]

[0032] In the formula, φ is the saturation function; T is the scale parameter of the saturation function. s λ is the discrete sampling time; N is the number of discrete terms; λ∈[λ min ,λ max ] represents the synovial gain; η∈[η min ,η max ] represents the sliding surface scaling factor; β∈[β min ,β max ] represents the fractional order; α∈[α min ,α max [e] represents the power exponent parameter of the control law; T This represents thrust error.

[0033] Preferably, in step S4), the zero-speed current suppression module ZSCR limiting strategy is as follows:

[0034]

[0035] in:

[0036]

[0037] In the formula, u final (t) represents the output of the final thrust loop control; v(t) represents the mechanical speed of the motor; v th The threshold for zero speed determination; u lim This is the maximum output value.

[0038] The beneficial effects of this invention are as follows:

[0039] 1. The hybrid observation strategy of this invention takes into account both high and low speed conditions. The low-speed ST-SMC+EKF improves the noise resistance, while the high-speed Luenberger+PLL ensures real-time performance. The observation accuracy is improved by more than 30%, which can adapt to the complex requirements of dual-use water and air scenarios. At the same time, it ensures control continuity and enhances the robustness of the system under complex water and air conditions.

[0040] 2. This invention, through a smooth switching mechanism based on dynamic hysteresis, can ensure a stable transition of the system when the speed of the sensorless motor is at the critical state, eliminate high-low speed switching jitter, and greatly reduce the error of critical point angular velocity and angle; the fluctuation of angle and speed estimation is reduced by more than 40%.

[0041] 3. The sensorless motor direct thrust control method designed in this invention for different water and air environments can better follow the thrust required by the flight control after calculation, compared with the traditional RPM speed control, and eliminate the error caused by the traditional speed-to-thrust mapping. This invention can also correct thrust deviation in a timely manner, improve the accuracy of flight control response, and optimize control precision; the thrust response speed is improved by more than 25%, and the control error is reduced to less than 5%.

[0042] 4. This invention enhances the system's adaptability to changes in water-air media through fuzzy-adaptive fractional sliding mode control, significantly improving robustness. Moreover, it can estimate thrust even in water-air cross-media environments, reducing the complexity of the control link and improving system reliability. Attached Figure Description

[0043] Figure 1 This is a schematic flowchart of the method of the present invention;

[0044] Figure 2 This is a control framework diagram of the method of the present invention;

[0045] Figure 3 This is a schematic diagram of the air thrust ring control process of the present invention;

[0046] Figure 4 This is a schematic diagram of the underwater thrust ring control process of the present invention. Detailed Implementation

[0047] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings:

[0048] like Figure 1 and 2 As shown, this embodiment provides a direct thrust control method for a dual-purpose (water and air) sensorless motor based on a hybrid observation strategy, including the following steps:

[0049] S1) A rotor position and speed observation strategy using speed-segment sensorless hybrid control is adopted, dynamically selecting low-speed and medium-to-high-speed observation algorithms based on the motor's mechanical speed; when the motor is in the air and operating at medium-to-high speed, a medium-to-high-speed observation algorithm combining the Luenberger state observation algorithm and the phase-locked loop (PLL) is adopted; when the motor is in the water and operating at low speed, a low-speed observation algorithm combining the super torsional sliding mode algorithm and the Kalman filter algorithm is adopted.

[0050] The low-speed observation algorithm, which combines the super torsional sliding mode algorithm and the Kalman filter algorithm, is used to estimate the electric angle and electric angular velocity of the motor rotor under low-speed conditions in water. Specifically, the algorithm includes the following steps:

[0051] S11) The sliding mode control terms are generated using a super torsional sliding mode algorithm; the details are as follows:

[0052] S111) A nonlinear switching function is introduced into the super torsional sliding mode observer to enhance robustness to parameter variations and noise, and the motor voltage equation is established in the α-β coordinate system based on the characteristics of the motor:

[0053]

[0054] In the formula, V a V β These represent the voltages of the motor along the α and β axes, respectively; I α I β E represents the motor current along the α and β axes, respectively. α E β These represent the back electromotive forces of the motor along the α and β axes, respectively; R s For stator resistance; L s For stator inductance;

[0055] S112) Construct a mathematical model for estimating the system state:

[0056]

[0057] In the formula, These are the estimated currents of the motor along the α and β axes, respectively. The back electromotive force of the motor is estimated along the α and β axes; u α u β These are the sliding mode control terms on the α and β axes, respectively;

[0058] S113) Define the sliding surface as the current error:

[0059]

[0060] In the formula, S α S βThese represent the current errors along the α and β axes, respectively.

[0061] S114) Generating sliding mode control terms through a super torsional control law, wherein the control law satisfies:

[0062]

[0063] In the formula, k1 and k2 are control gains; v α v β These are auxiliary state variables; These represent the rates of change of the auxiliary state variables; sign represents the sign function.

[0064] When S α →0 and When the sliding surface reaches the sliding mode, the following applies:

[0065]

[0066] Introducing gain k e ,get:

[0067]

[0068] Ideally, when the gain k e When = 1, we have:

[0069]

[0070] S12) Use the extended Kalman filter algorithm to predict and update the optimized electrical angle estimate. and electric angular velocity estimation

[0071] S121) Define the state variable x as:

[0072]

[0073] In the formula, θ e ω is the electrical angle; e Electric angular velocity;

[0074] S122) Based on the state variable x, the state equation is constructed as follows:

[0075]

[0076] In the formula, The time derivative of the state vector; Electric angular velocity; Indicates electric angular acceleration;

[0077] f(x) represents the state transition function; w is the process noise;

[0078] S123), the measurement equation is constructed as follows:

[0079] z = h(x) + v; (10)

[0080] in:

[0081]

[0082] In the formula, v is the measurement noise; τ is the back electromotive force coefficient; h(x) is the mapping function from state to measurement; and z is the back electromotive force matrix.

[0083] In this embodiment, the Luenberger observation algorithm and a phase-locked loop (PLL) are used to estimate the electrical angle and electrical angular velocity of the motor rotor when it is in the air at medium to high speeds; specifically, the following steps are included:

[0084] S131) ​​Construct a state equation containing current and back electromotive force, and achieve state estimation through the observer gain matrix, i.e.:

[0085]

[0086] In the formula, A is the system state matrix; B is the system input matrix; and G is the gain matrix of the Luenberger observer.

[0087] S132) Using a phase-locked loop (PLL) to obtain the estimated back electromotive force The estimated electrical angle and electrical angular velocity are extracted from the calculated back electromotive force. The phase-locked loop feedback control loop aligns the estimated angle with the actual angle. The phase error is calculated, and then the estimated electrical angle and electrical angular velocity are adjusted by a PI controller. That is:

[0088]

[0089] In the formula, Δe is the back electromotive force phase error; To estimate the electrical angle; To estimate the electric angular velocity; K p K is the proportional coefficient of the controller. i For controller integral coefficients; phase error

[0090] The relationship between back electromotive force and electric angle is as follows:

[0091]

[0092] In the formula, τ is the back electromotive force coefficient; θ e ω is the electrical angle; e ω is the electric angular velocity.

[0093] S2) A smooth switching mechanism based on dynamic hysteresis is used to switch between high-speed and low-speed observation algorithms, and a weighted fusion algorithm is used to achieve smooth fusion of the outputs of the two algorithms within the dynamic hysteresis interval; specifically, the following steps are included:

[0094] S21), Set the hysteresis interval [ω] low ,ω high In this embodiment, ω low =450RPM, ω high =550 RPM;

[0095] S22) Dynamically adjust the hysteresis boundary according to the velocity change trend, that is:

[0096]

[0097] In the formula, ω′ low ,ω′ hihh The adjusted hysteresis boundary; k d For dynamic adjustment coefficients; Δω m Let Δω be the change in velocity. m =ω m (k)-ω m (k-1), ω m (k) represents the motor speed at the k-th sampling time;

[0098] S23) Switch the observation strategy according to the following conditions:

[0099] a. If the current algorithm is a super torsional sliding mode algorithm and a Kalman filter algorithm, and ω m >ω ′ high Then switch to the Luenberger observation algorithm and the phase-locked loop (PLL) algorithm;

[0100] b. If the current algorithm is Luenberger observation algorithm and phase-locked loop (PLL) algorithm, and ω m <ω′ low Switch to the super torsional sliding mode algorithm and the Kalman filter algorithm;

[0101] c. If the hysteresis interval is within the range, apply an exponentially decaying nonlinear weighted fusion algorithm to fuse the outputs of the two algorithms.

[0102] In this embodiment, if the outputs of the two algorithms are fused using a nonlinear weighted fusion algorithm based on exponential decay within the hysteresis interval, then:

[0103]

[0104] In the formula, To estimate the electrical angle; To estimate the electric angular velocity; For fusion factor; This represents the midpoint of the hysteresis loop; δ is the attenuation coefficient. These correspond to the estimated electrical angle and estimated electrical angular velocity obtained by the super torsional sliding mode algorithm and the Kalman filter algorithm, respectively. The estimated electrical angle and estimated electrical angular velocity are obtained by the Luenberger observation algorithm and the phase-locked loop (PLL) algorithm, respectively.

[0105] S3) Direct thrust control models are constructed for water-air cross-medium scenarios. In the air, a thrust formula based on Bernoulli's equation, momentum theorem and leaf element theory is used. In the water, a correction factor for the influence of the underwater environment is introduced to adjust the direct thrust in the air to obtain the direct thrust of the motor in the water.

[0106] The propeller thrust T in mid-air is calculated using Bernoulli's theorem and leaf element theory; that is:

[0107]

[0108] In the formula, T is the propeller thrust in the air, and ρ air R is the air density; R is the propeller radius; v i For in-flight induced velocity; C is the average chord length of the propeller blades; L ω is the lift coefficient; ω is the angular velocity of the propeller.

[0109] The aerial propulsion efficiency k is calculated based on aerodynamic power and electromechanical power; that is:

[0110]

[0111] In the formula, k is the aerial propulsion efficiency; p is the number of motor pole pairs; ψ is the permanent magnet flux linkage; I q This is the q-axis current;

[0112] Finally, the direct aerial thrust T of the motor is calculated based on the propeller thrust and aerial propulsion efficiency. air ;Right now:

[0113]

[0114] In the formula, T air This is the direct thrust of the motor in mid-air.

[0115] Correction factor K for underwater environmental impacts introduced into water env By correction factor K env Direct thrust T in the air air Adjustments were made to obtain the direct thrust T of the motor in water. water ,Right now:

[0116]

[0117] in:

[0118]

[0119] In the formula, k ′ For underwater propulsion efficiency; K ρ K is the density influence coefficient. k ρ is the efficiency loss coefficient. air ρ is the density of air. water θ represents the water density; θ is the efficiency compensation factor.

[0120] S4) Direct thrust is controlled using a fuzzy-adaptive fractional-order sliding mode controller. The fuzzy-adaptive fractional-order sliding mode controller includes a fuzzy inference module, a fractional-order sliding mode controller, and a zero-speed current suppression module (ZSCR). The fuzzy inference module dynamically adjusts the sliding mode controller parameters based on the current system mode and error amplitude. The fractional-order sliding mode controller constructs the sliding surface s(t) and control law u(t) based on the output parameters λ, η, β, and α of the fuzzy inference module. The zero-speed current suppression module (ZSCR) implements a limiting strategy on the output control quantity u(t). Specifically, the following steps are included:

[0121] S41) Define thrust error e T for:

[0122] e T =T d -T now ; (twenty two)

[0123] In the formula, T d Thrust is controlled to target; T now For current thrust;

[0124]

[0125] S42), based on thrust error e T The current motor mode is given to the fuzzy inference module, which outputs the dynamic parameters of the adaptive fractional-order sliding diaphragm controller.

[0126] Among them: the motor mode includes air thrust mode and underwater thrust mode;

[0127] The dynamic parameters of the fractional-order slug controller include the slug gain λ∈[λ]. min ,λ max ]、Slip surface scaling factor η∈[η min ,η macx ], fractional order β∈[β min ,β max ], the power exponent parameter of the control law α∈[α min,α max ].

[0128] The specific dynamic output parameters are calculated based on the fuzzy rule base, as shown in Table 1:

[0129] Table 1 Fuzzy Rule Table

[0130]

[0131] S43) The λ, η, β, and α output by the fuzzy inference module are directly used to construct the sliding surface s(t) and the control law u(t) in the fractional-order sliding controller, that is:

[0132]

[0133] in:

[0134]

[0135] In the formula, φ is the saturation function; T is the scale parameter of the saturation function. s Where N is the discrete sampling time; N is the number of discrete terms;

[0136] S44) To prevent current surges caused by sudden changes in control signals during the zero-speed or low-speed startup phase of the system, a zero-speed current suppression module (ZSCR) is designed. The limiting strategy of the ZSCR is as follows:

[0137]

[0138] in:

[0139]

[0140] In the formula, u final (t) represents the output of the final thrust loop control; v(t) represents the mechanical speed of the motor; v th The threshold for zero speed determination; u lim This is the maximum output value.

[0141] The control framework diagrams for thrust in the air and underwater are shown below. Figure 3 , 4 As shown.

[0142] The embodiments and descriptions above are merely illustrative of the principles and preferred embodiments of the present invention. Various changes and modifications may be made to the present invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed.

Claims

1. A direct thrust control method for a dual-purpose (water and air) sensorless motor based on a hybrid observation strategy, characterized in that, Includes the following steps: S1) A rotor position and speed observation strategy using speed-segment sensorless hybrid control is adopted, and the low-speed observation algorithm and the medium- and high-speed observation algorithm are dynamically selected according to the motor mechanical speed to estimate the motor rotor electrical angle and electrical angular velocity. S2) Based on the smooth switching mechanism of dynamic hysteresis, the high-speed and low-speed observation algorithm is switched, and the output of the two algorithms within the dynamic hysteresis interval is smoothly fused through the weighted fusion algorithm. S3) Direct thrust control models are constructed for water-air cross-medium scenarios. In the air, a thrust formula based on Bernoulli's equation, momentum theorem and leaf element theory is used. In the water, a correction factor for the influence of the underwater environment is introduced to adjust the direct thrust in the air to obtain the direct thrust of the motor in the water. S4) Direct thrust is controlled using a fuzzy-adaptive fractional sliding membrane controller.

2. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy according to claim 1, characterized in that: In step S1), when the motor is in the air at medium to high speed, a medium to high speed observation algorithm combining the Luenberger state observation algorithm and the phase-locked loop (PLL) is used; when the motor is in the water at low speed, a low speed observation algorithm combining the super torsional sliding mode algorithm and the Kalman filter algorithm is used.

3. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy, as described in claim 2, is characterized in that: In step S1), a low-speed observation algorithm combining the super torsional sliding mode algorithm and the Kalman filter algorithm is used to estimate the electrical angle and electrical angular velocity of the motor rotor, specifically as follows: Construct the motor voltage equation in the α-β coordinate system and define the sliding surface S. α S β For the current error, a sliding mode control term is generated using a super torsional control law, where the control law satisfies: In the formula, k1 and k2 are control gains; v α v β These are auxiliary state variables; These represent the rates of change of the auxiliary state variables; sign represents the sign function. The extended Kalman filter algorithm uses the electric angle and electric angular velocity of the motor as state variables. It optimizes the back EMF observations through state equations and measurement equations to achieve accurate estimation of rotor position and speed.

4. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy, as described in claim 3, is characterized in that: In step S2), a dynamic hysteresis interval is introduced, and the hysteresis boundary is dynamically adjusted according to the trend of velocity increase or decrease. Within the hysteresis interval, the outputs of the two algorithms are fused using a nonlinear weighted fusion algorithm based on exponential decay.

5. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy, as described in claim 4, is characterized in that: In step S2), the outputs of the two algorithms are fused using a nonlinear weighted fusion algorithm based on exponential decay, specifically as follows: In the formula, To estimate the electrical angle; To estimate the electric angular velocity; For fusion factor; Indicates the midpoint of the hysteresis loop; δ is the attenuation coefficient; ω′ low ,ω′ high The adjusted hysteresis boundary; ω m (k) represents the motor speed at the k-th sampling time; These correspond to the estimated electrical angle and estimated electrical angular velocity obtained by the super torsional sliding mode algorithm and the Kalman filter algorithm, respectively. The estimated electrical angle and estimated electrical angular velocity are obtained by the Luenberger observation algorithm and the phase-locked loop (PLL) algorithm, respectively.

6. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy, as described in claim 5, is characterized in that: In step S3), the direct air thrust Y of the motor air for: In the formula, T air ψ is the direct thrust of the motor in the air; k is the in-air propulsion efficiency; p is the number of pole pairs of the motor; ψ is the flux linkage of the permanent magnet; I q ω is the q-axis current; ω is the angular velocity of the propeller. C is the average chord length of the propeller blades; L This is the lift coefficient.

7. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy according to claim 6, characterized in that: In step S3), the direct thrust T of the motor in the water water for: In the formula, K env is the correction factor; k′ is the underwater propulsion efficiency.

8. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy according to claim 1, characterized in that: In step S4), the fuzzy-adaptive fractional-order sliding mode controller includes a fuzzy inference module, a fractional-order sliding mode controller, and a zero-speed current suppression module ZSCR. The fuzzy inference module dynamically adjusts the sliding mode controller parameters according to the current system mode and error amplitude. The fractional-order sliding mode controller constructs the sliding surface s(t) and control law u(t) according to the output parameters λ, η, β, and α of the fuzzy inference module. The zero-speed current suppression module ZSCR implements a limiting strategy on the output control quantity u(t).

9. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy, as described in claim 8, is characterized in that: In step S4), the fractional-order sliding mode controller constructs the sliding surface s(t) and the control law u(t) based on the output parameters λ, η, β, and α of the fuzzy inference module, specifically as follows: In the formula, φ is the saturation function; φ is the scale parameter of the saturation function. T s λ is the discrete sampling time; N is the number of discrete terms; λ∈[λ min ,λ max ] represents the synovial gain; η∈[η min ,η max [ ] represents the sliding surface scaling factor; β∈[β min ,β max ] represents the fractional order; α∈[α min ,α max [e] represents the power exponent parameter of the control law; T This represents thrust error.

10. The method for direct thrust control of a sensorless motor for both water and air applications based on a hybrid observation strategy, as described in claim 9, is characterized in that: In step S4), the zero-speed current suppression module ZSCR limiting strategy is as follows: in: In the formula, u final (t) represents the output of the final thrust loop control; v(t) represents the mechanical speed of the motor; v th The threshold for zero speed determination; u lim This is the maximum output value.

Images