De-icing spray control method and system for a rotating blade de-icing robot

Abstract

This invention relates to the field of wind turbine maintenance, specifically to a de-icing jet control method and system for a rotating blade de-icing robot. The control method acquires the comprehensive external load vector in a rotating reference frame in real time and estimates the transient recoil vector of the pulsed jet, calculating the jet phase window that meets the fuselage adhesion torque margin requirements. During the de-icing operation within this window, the control fluid distribution component synchronously guides the high-pressure fluid to the first injection end for de-icing and to the second injection end, thereby generating transient aerodynamic downpressure within the omnidirectional convergent surface of the fuselage belly, achieving a dynamic balance between jet flow and recoil. This solution, through the coordinated control of aerodynamic stabilization and mechanical compensation, effectively mitigates the nose-up, overturning, and yaw tendencies caused by high-pressure pulses, significantly improving the adhesion reliability and operational accuracy of the lightweight robot under rotating variable load and variable curvature blade environments, without requiring additional fuselage mass.

Description

Technical Field

[0001] This invention relates to wind turbine maintenance, and more specifically to a de-icing jet control method and system for a rotary blade de-icing robot. Background Technology

[0002] As wind power technology advances towards larger megawatts and ultra-long blades, the need for anti-icing and de-icing maintenance of wind turbines in freezing winter environments is becoming increasingly prominent. Traditional shutdown de-icing methods result in significant economic losses; therefore, using wall-climbing robots for online or low-speed blade de-icing has become an industry trend.

[0003] Therefore, in the wall-climbing de-icing robot technology, high-pressure pulsed airflow (or pneumatic and ultrasonic / electrothermal coupling) is used to break up and clean the hard ice in order to effectively remove it. At the same time, in order to reduce the huge centrifugal force and gravitational alternating load on the robot body when it rotates with the blade, the applicant has developed a de-icing robot solution, which adopts a mother-daughter / base station lightweight architecture. That is, the heavy power and air source equipment (such as air compressor and battery pack) is externally placed on the wind turbine hub base station through an umbilical cable, so that the end of the robot climbing on the blade can be simplified and lightweight.

[0004] However, regardless of whether it is a non-mother-daughter type robot or a lightweight architecture type robot, there is a significant dynamic mechanical conflict between them and high-energy pulse de-icing:

[0005] When a robot's main nozzle sprays a high-pressure pulse of air into the ice ahead, according to the law of conservation of momentum, a huge backward aerodynamic reaction force (i.e., recoil) is instantly generated. For traditional robots that independently carry heavy components, their large body mass can still reduce the impact by their own inertia (but it still has an adverse effect on their grip); for lightweight base station robots, their smaller body mass (i.e., smaller inertial mass) is under even greater threat, and the situation of damping such millisecond-level transient high-energy impacts is even more severe.

[0006] Even more serious is that when the robot is in the reference frame of the rotating blade, the recoil force of the pulse not only generates a backward translational thrust, but also forms a torque arm with the robot's supporting center of gravity, causing a strong upward lifting tendency of the front of the fuselage. Once the front of the fuselage is lifted instantaneously, the vacuum suction cups that were originally in close contact with the blade surface will leak air at the edges, and the original aerodynamic flow channels of the chassis will also be damaged, resulting in a precipitous decrease in negative pressure adhesion. At the same time, along with the Coriolis force and directional gravity generated by the blade rotation, this unpredictable transient coupling torque can easily cause the lightweight robot to tilt, slide laterally, or even detach directly from the blade at a high altitude.

[0007] Existing control schemes for climbing suction cup robots rely solely on passively increasing the static suction force of the suction cups. This not only results in a severe lag in response to millisecond-level pulse impacts, but also blindly increases the suction force, greatly increasing the system's energy consumption and wear. Furthermore, it cannot effectively reduce the transient instability traps caused by pulse recoil. Summary of the Invention

[0008] To address the aforementioned technical problems, the present invention aims to provide a control method and system for a rotating blade de-icing robot. This system aims to solve the problem that existing de-icing robots, when operating under conditions of online wind turbine operation or low-speed rotation, are prone to tipping over, lateral slippage, and adhesion failure due to the coupling of the transient recoil vector generated by high-pressure pulse jet with the combined loads of gravity, centrifugal force, and Coriolis force in the rotating reference frame.

[0009] This invention constructs a pulse triggering mechanism based on force prediction and phase locking, combined with multi-path transient diversion technology of fluid kinetic energy, to transform the harmful recoil torque generated by de-icing into aerodynamic downforce and lateral correction torque that enhance the fit stability of the robot, thereby improving the high dynamic adhesion stability of the robot under the complex force field of changing direction and the working surface of changing curvature.

[0010] This invention provides a de-icing spray control method for a rotating blade de-icing robot. The method operates on a de-icing robot that moves on a rotating blade via an attachment unit formed by a single vacuum suction cup foot or multiple vacuum suction cup foot groups. The robot is equipped with a fluid distribution component. The input end of the fluid distribution component is connected to a high-pressure fluid source, and its output end branches at least into a first spray end pointing towards the de-icing operation path and a second spray end pointing towards the converging curved surface at the bottom of the robot body. The control method includes:

[0011] The system acquires the combined external load vector generated by the forces acting on the robot in the rotating reference frame in real time. The combined external load vector is composed of at least the gravity of the robot body, the aerodynamic forces of the environment, and the rotational inertial force. Based on the parameters of the de-icing jet to be executed, the system estimates the transient recoil vector generated at the moment of triggering the first jet.

[0012] Based on the spatial vector relationship between the integrated external load vector and the transient recoil vector, the jet phase window that satisfies the preset threshold for the fuselage adhesion torque margin is calculated.

[0013] If it is determined that the current jet phase window is in effect and de-icing operation is triggered, the control fluid distribution component performs transient multi-channel fluid distribution: the first proportion of high-pressure fluid is directed to the first jet end to output the ice-breaking jet, and the second proportion of high-pressure fluid is simultaneously directed to the second jet end to drive the high-pressure fluid to sweep tangentially along the convergent curved surface at the bottom of the fuselage, so as to generate transient aerodynamic downpressure pointing towards the working surface.

[0014] The aforementioned comprehensive external load vector is the three-dimensional vector resultant force of all external physical forces acting on the entire de-icing robot within the non-inertial reference frame of the rotating fan. The aforementioned fuselage adhesion torque margin is a quantitative indicator of the robot's adhesion stability, representing the difference between the stabilizing clamping torque and the instability overturning torque acting on the fuselage.

[0015] When the robot performs de-icing spraying, the jet generates a recoil and overturning moment that attempts to tip the robot over. The robot can remain stable only if the existing clamping torque (including the torque generated by the suction cups and the torque pointing towards the blades in the combined external load) has a remainder after deducting the recoil moment; this remainder is the margin.

[0016] This solution utilizes the combined force of transient aerodynamic downforce and the comprehensive external load vector to construct a constraint torque that dynamically counteracts the transient recoil vector, thereby mitigating the tendency of the fuselage to pitch up or yaw due to forced tilting. Addressing the technical challenge of de-icing robots, especially lightweight de-icing robots operating under rotating blade conditions, where the coupling of transient high-energy jet recoil force and centrifugal load can easily lead to nose-up tilting and adhesion failure, this solution cleverly achieves dynamic counterbalancing by using the jet to counteract the recoil. This transforms the harmful recoil load, which would otherwise destabilize adhesion, into a beneficial clamping torque that enhances fuselage adhesion. This not only reduces the risk of transient nose-up tilting from its physical source but also significantly improves the robot's damping capacity against recoil impacts without adding extra fuselage mass, greatly enhancing the adhesion reliability of lightweight robots under extreme directional composite load environments.

[0017] In some embodiments, the jet phase window for which the fuselage adhesion torque margin satisfies a preset threshold is calculated based on the spatial vector relationship between the integrated external load vector and the transient recoil vector, including the following steps:

[0018] When the transient recoil vector acts on the fuselage, the potential overturning axis that causes the fuselage to tilt relative to the working surface is determined. Using the potential overturning axis as a reference, the sequence of comprehensive external load torques generated by the dynamic mapping of the comprehensive external load vector in the 360-degree full phase domain of the fan rotor rotation is calculated, as well as the recoil overturning torque generated by the transient recoil vector relative to the potential overturning axis is calculated. The comprehensive external load torque sequence is traversed to select the phase intervals that make the comprehensive external load torque manifest as a stable clamping torque pointing towards the working surface. Within the phase interval, if the sum of the stable clamping torque and the effective adsorption torque provided by the current robot's attachment unit, minus the recoil overturning torque, results in a residual constraint torque greater than a preset safety reserve threshold, then the corresponding continuous phase interval is marked as the jet phase window.

[0019] The aforementioned stabilizing clamping torque is the resultant torque component generated by the synthesis of gravity vector, rotational inertial force (including centrifugal force and Coriolis force), and environmental aerodynamic force vector in the rotating reference frame in which the wind turbine operates, and its spatial orientation is towards the working surface (i.e., the surface of the wind turbine blade). When this torque acts on the fuselage, it can suppress the tendency of the fuselage to pitch up, tilt back, or detach from the working surface, thereby jointly constructing the stable adhesion state of the fuselage with the active adsorption force generated by the adhesion unit.

[0020] By determining the potential overturning axis and using it as a benchmark to calculate the comprehensive external load torque sequence across the entire phase domain, this scheme can utilize the natural load trends during wind turbine rotation. This scheme can accurately identify the golden operating phase where the resultant force of gravity and rotational inertia is in a natural compressive state. By establishing a full-dimensional torque discrimination criterion including adsorption torque, background load torque, and recoil overturning torque, it ensures that high-pressure pulses are triggered only within a window period when the system has sufficient adhesion margin. The control logic of this scheme effectively avoids the instability risk caused by the superposition of natural loads and jet recoil in the time dimension, providing a solid logical prediction guarantee for the safe operation of the aforementioned robots, especially lightweight robots, under extremely variable force fields.

[0021] In some embodiments, the step of controlling the fluid distribution component to perform transient multi-channel fluid distribution and dynamically adjusting the first ratio and the second ratio includes: obtaining the target pressure of the main jet required for the current de-icing operation, and forward calculating the transient recoil torque generated at the moment of triggering based on the target pressure of the main jet and the nozzle cross-sectional area of ​​the first injection end; calculating the target aerodynamic downforce required to counteract the transient recoil torque based on the torque balance constraint condition; calling the preset fluid dynamic acceleration mapping model of the convergent surface of the fuselage bottom, and backward calculating the tangential fluid target velocity required to generate the target aerodynamic downforce within the convergent surface; calculating the dynamic distribution commands of the first ratio and the second ratio according to the fluid consumption corresponding to the target pressure of the main jet and the fluid consumption corresponding to the tangential fluid target velocity, and outputting them to the drive control terminal of the fluid distribution component; wherein, the greater the target pressure of the main jet, the greater the fluid kinetic energy of the second ratio controlled to be distributed to the second injection end, so as to make the transient aerodynamic downforce generated at the fuselage bottom increase with the intensity of the ice-breaking jet at the first injection end.

[0022] The forward calculation described above, based on known input conditions (nozzle size, preset de-icing pressure), uses Newton's third law and the momentum theorem to calculate the required destructive force (recoil torque), thus predicting the magnitude of the disturbance. The reverse calculation, based on known objectives (targeted aerodynamic pressure, i.e., the force required to hold down the robot), substitutes this target value into the fluid dynamics model (Venturi acceleration model) to calculate the required airflow speed from the abdominal nozzle to generate the necessary suction, thus solving for the countermeasure parameters. This scheme utilizes forward calculation to predict the destructive torque generated by the main jet and, combined with reverse calculation, dynamically solves for the compensation velocity required for the abdominal convergent surface, achieving a high degree of alignment between control parameters and physical requirements. The higher the blowing pressure of the main nozzle, the stronger the synchronous transient pressure on the abdomen it generates, thus constructing a self-suppressing system at the physical level. This ensures that the fuselage can maintain a stable stress state when facing pulse impacts of different energy levels, reducing the risk of fuselage pulsation or rollover caused by jet recoil.

[0023] In some embodiments, the control method further includes an adaptive closed-loop correction step for transient aerodynamic downpressure:

[0024] The system monitors the adhesion status parameters fed back by each adhesion unit in real time, as well as the displacement deviation value of the fuselage relative to the working surface. It extracts adhesion performance loss characteristics from the adhesion status parameters or displacement deviation values ​​and couples them with the targeted aerodynamic downpressure to extract downpressure compensation deviation values. An adaptive correction factor is generated based on the downpressure compensation deviation value, and this factor is used to reconstruct the hydrodynamic acceleration mapping model in real time. The second ratio is dynamically updated according to the reconstructed hydrodynamic acceleration mapping model to compensate for downpressure conversion losses caused by changes in working surface roughness or fuselage attitude deflection. This closed-loop correction mechanism allows for real-time sensing of actual aerodynamic pressure fluctuations caused by uneven working surface roughness, residual ice interference, or slight fuselage attitude deflection, and provides millisecond-level dynamic compensation for the flow splitting ratio through the adaptive correction factor. This closed-loop feedback control significantly improves the system's robustness, ensuring that even on non-ideal contact or irregular working surfaces, the fuselage bottom can still generate the expected targeted aerodynamic downpressure, achieving high reliability in maintaining the adhesion status.

[0025] In some embodiments, the control method further includes a dynamic compensation step for the atmospheric physical environment, including:

[0026] The system acquires real-time atmospheric pressure and ambient temperature of the operating environment and calculates air density correction coefficients under current altitude and meteorological conditions. These correction coefficients are then used to calibrate the parameters of the fluid dynamics acceleration mapping model, dynamically correcting the mapping function between the tangential fluid target velocity and transient aerodynamic downpressure based on changes in the atmospheric physical environment. The second proportional gain is updated based on the corrected mapping function, enabling precise control of transient aerodynamic downpressure on the transient recoil vector under different atmospheric density conditions. Since the aerodynamic downpressure generated by the Venturi effect is highly dependent on air density, this solution can automatically adjust the fluid distribution strategy according to altitude (e.g., mountain wind fields versus ocean wind fields) and climate change, reducing undercompensation or overcompensation caused by atmospheric environmental changes. This technology ensures that the robot maintains a constant and accurate load offset capability at different latitudes, longitudes, and altitudes globally, significantly expanding the system's operational boundaries.

[0027] In some embodiments, the output end of the fluid distribution component further branches into third jet ends pointing to the attachment units on the left and right sides of the fuselage, respectively; the control method also includes a lateral correction step for suppressing yaw slip, including: acquiring the rotational angular velocity of the fan rotor and the displacement velocity of the robot in real time, calculating the Coriolis force component generated by their coupling, and combining the lateral eccentricity of the transient recoil vector to synthesize the transient yaw disturbance torque; according to the spatial deflection direction and amplitude of the transient yaw disturbance torque, controlling the fluid distribution component to perform asymmetric differential fluid distribution to the third jet ends located on the left and right sides of the fuselage; controlling the third jet end on the same side as the deflection trend of the transient yaw disturbance torque to instantaneously increase the flow split ratio and output a lateral pressurized jet, so as to use the recoil momentum generated by the lateral pressurized jet to construct a lateral correction torque that dynamically counterbalances the transient yaw disturbance torque; simultaneously using the lateral pressurized jet to blow away the derived water film and residual ice between the attachment unit on this side and the working surface, so as to improve the anti-slip static friction boundary of the attachment interface. By calculating the Coriolis force component and coordinating with lateral adjustment nozzles to execute asymmetric differential jets, the corrective torque generated by the asymmetric flow splitting can precisely counteract the lateral deviation caused by rotational inertia, maintaining the straightness of the robot's trajectory. Simultaneously, the lateral airflow synchronously sweeps the adsorption interface, eliminating lubrication issues caused by ice fragments and derived water films, thus physically enhancing the static friction boundary of the attachment unit. This dual guarantee of active thrust correction and enhanced interface friction ensures that the robot maintains extremely high trajectory-keeping and anti-slip capabilities even under complex directional force fields.

[0028] In some embodiments, the control method further includes a coordinated interlocking step between the jet and the gait, specifically comprising:

[0029] The robot monitors the operating phase of each attachment unit in real time to determine whether the robot body is currently in a stepping phase where some attachment units are detached from the working surface, or in a supporting phase where all attachment units are attached to the working surface. An interlocking mechanism is established between the jet trigger command for de-icing operations and the operating phase: if the robot body is currently in a stepping phase, the jet trigger command is suspended, restricting the fluid distribution component from performing transient multi-channel fluid distribution until the robot body returns to the supporting phase; if the robot body is currently in the supporting phase and de-icing operations are triggered, during the synchronization of the fluid distribution component performing transient multi-channel fluid distribution, the drive joints of each attachment unit are controlled to enter a rigid locking mode; a rigid transmission path is constructed using the attachment units in rigid locking mode to penetrate the robot body and release the impact load caused by the transient recoil vector to the working surface, thereby suppressing the robot body resonance and attitude instability caused by the flexible displacement of the joints. The aforementioned rigid locking mode refers to a high-rigidity suppression state of the joints set up to resist the impact of high-energy jet recoil during the transient period triggered by the de-icing pulse. It aims to reduce the flexible gap and elastic damping of the drive joints, so that the robot as a whole forms a temporary, non-deformable rigid truss at the physical level.

[0030] By establishing an interlocking mechanism between the operating phase and the jet command, and a rigid locking mode for the drive joints, this solution mitigates the risk of instability during the robot's most vulnerable moments of movement. Forcibly suspending the jet command during the stepping phase prevents the fuselage from bearing impact when the number of attached units decreases; while the rigid locking of the support phase reduces the flexibility and elastic resonance of the mechanical joints, creating an "energy dissipation path" that directly guides the recoil impact to the large blade substrate. This solution significantly suppresses fuselage flutter caused by pulsed loads, protects the precision actuators, and ensures perfect coordination in timing and stiffness between high-energy ice-breaking and climbing displacement actions.

[0031] In some embodiments, the control method further includes a dynamic protection step based on the coupling of thermal load and aerodynamic pressure, including: real-time acquisition of the ambient temperature and system status, the system status including at least the drive joint resistance and edge sealing margin of each attachment unit; if the ambient temperature is determined to be lower than a preset temperature threshold, the fluid distribution component guides the thermo-pressurized fluid generated by the high-pressure fluid source to perform anti-icing heat maintenance operation: a set proportion of thermo-pressurized fluid is directed to the edge guide gaps of the drive joint and attachment unit; this is used to suppress joint freezing resistance by utilizing heat conduction and simultaneously reduce the risk of frost and leakage at the edge of the attachment unit by utilizing thermal convection; if the comprehensive external load vector is determined to reach the instability boundary during the de-icing operation, an energy redistribution action is performed: a forced switch to a second distribution ratio is made, and the full amount of thermo-pressurized fluid is directed to the second injection end, so as to generate the maximum transient aerodynamic downpressure by utilizing the full fluid kinetic energy. By combining the fluid distribution component with the thermo-pressurized fluid, this solution achieves a deep integration of operational efficiency and system survivability. Utilizing the thermal energy of compressed air for heat maintenance aims to solve the problems of drive joint freezing causing motion lag and suction cup edge frost causing sealing failure in cold environments. The full redistribution action at the instability boundary ensures that the maximum peak aerodynamic downpressure can be instantaneously generated under extreme sudden loads. This solution achieves a comprehensive protection effect of "normal antifreeze to maintain accuracy and critical pressurization to ensure survival" through coupled control of thermal and pressure energy fields, significantly improving the robot's continuous operation capability in extremely cold environments.

[0032] A second aspect of the present invention also provides a de-icing spraying system for a rotary blade de-icing robot. The de-icing spraying system includes: a fluid control hub, a robot chassis, a nozzle array, a main working nozzle, a negative pressure induction nozzle, a lateral adjustment nozzle, and a central processing unit. The fluid control hub includes a high-pressure fluid inlet and a multi-way proportional directional valve assembly connected to the high-pressure fluid inlet; the underside of the robot chassis is provided with an omnidirectional convergent surface, which smoothly bulges inward from the peripheral edge of the robot chassis, and the geometric extreme point of the omnidirectional convergent surface is located in the central region of the underside of the chassis; the nozzle array is distributed at different phases of the chassis, including the main working nozzle, the negative pressure induction nozzle, and the lateral adjustment nozzle. The main working nozzle is located at the front of the robot body, pointing towards the working path through the first output of the multi-way proportional directional valve group. The negative pressure induction nozzle is located in the front area of ​​the robot body, with its nozzle pointing towards the geometric extreme point of the omnidirectional convergent surface, and its spray axis is parallel to the tangent of the omnidirectional convergent surface on the spray path. Lateral adjustment nozzles are symmetrically arranged on the left and right sides of the robot body, pointing towards the attachment units on both sides. The two lateral adjustment nozzles are supplied with air through two independent third outputs of the multi-way proportional directional valve group. The central processing unit is electrically connected to the multi-way proportional directional valve group and is used to execute the aforementioned de-icing spray control method. The different phases mentioned above refer to the specific geometric orientation and angular distribution of each nozzle relative to the geometric center or longitudinal axis of the robot body in a three-dimensional spatial coordinate system (especially within the horizontal projection plane).

[0033] By incorporating a smoothly convex omnidirectional curved surface that rises inward from the periphery of the robot's chassis, and using nozzles that spray along the tangential direction, a highly efficient and symmetrical aerodynamic acceleration channel is constructed at the bottom of the fuselage. This configuration maximizes the activation of the Venturi effect and Bernoulli negative pressure, rapidly converting the kinetic energy of the high-pressure fluid diverted to the fuselage's underside into adhesion force perpendicularly directed towards the working surface. Simultaneously, the integration of a multi-way proportional reversing valve assembly enables precise on-demand allocation of the flow rates of the main, auxiliary, and lateral nozzles, ensuring a real-time balance between de-icing efficiency and body stability load. Furthermore, the symmetrically arranged lateral adjustment nozzles, combined with the control of the central processing unit, endow the system with the ability to correct asymmetric torques. This system addresses the shortcomings of insufficient impact inertia in lightweight fuselages from a physical architecture perspective, enhancing the de-icing robot's anti-tipping and anti-slip capabilities under extreme working conditions.

[0034] In some embodiments, the omnidirectional convergent surface is composed of a flexible covering layer laid on the bottom of the robot chassis; the inner cavity of the robot chassis is provided with a surface adjustment mechanism, which includes a displacement driving component and a top support member driven by the displacement driving component to perform lifting and lowering movements; the working end of the top support member abuts against the inner surface of the flexible covering layer, and is used to drive the flexible covering layer to deform through the lifting and lowering displacement of the top support member, so as to dynamically adjust the bulge height and curvature of the omnidirectional convergent surface; the surface adjustment mechanism also includes a locking component, which is used to fix or release the current lifting and lowering position of the top support member, so that the omnidirectional convergent surface can be adapted to the curvature of the current working surface.

[0035] In some embodiments, the inner cavity of the robot chassis is further provided with a dynamic center of gravity compensation mechanism, which includes a longitudinal slide rail arranged in the inner cavity and an active compensation mass block disposed on the longitudinal slide rail; the active compensation mass block is composed of functional components inside the body; during the synchronization of the release of the pulse jet at the first injection end, the active compensation mass block generates a forward compensation displacement toward the front end of the body, and uses the transient inertial torque generated by the forward compensation displacement to neutralize the pitching torque generated by the transient recoil vector on the body.

[0036] In some embodiments, the active compensation mass block is connected to the robot chassis via an elastic buffer component;

[0037] When the pulse jet is released at the first injection end, causing the fuselage to generate a backward instantaneous acceleration, the active compensation mass block uses physical inertia to compress the elastic buffer component and generate a forward compensation displacement.

[0038] A third aspect of the present invention also provides a rotary blade de-icing robot, including the de-icing spray system of the second aspect described above, an attachment drive component, and a power and medium supply component; the attachment drive component is symmetrically arrayed on both sides of the robot chassis, including multiple independently controlled vacuum adsorption feet, and each vacuum adsorption foot is provided with an active drive joint for performing step displacement, the active drive joint being electrically connected to a central processing unit; the power and medium supply component includes a supply base station set independently of the robot body, and an umbilical cable connecting the supply base station and the robot body; wherein, the supply base station integrates a high-pressure air source, and the umbilical cable integrates a high-pressure air supply hose, one end of the high-pressure air supply hose being connected to the high-pressure air source of the supply base station, and the other end being connected to a high-pressure fluid inlet on the robot body, thus achieving a lightweight robot body by externalizing the base station. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating the de-icing jet control method for a rotating blade de-icing robot in the embodiments.

[0040] Figure 2 This is a side view schematic diagram illustrating the de-icing robot in the implementation method;

[0041] Figure 3 This is a schematic diagram illustrating the double-layered abdominal structure of the de-icing robot in the embodiment.

[0042] Figure 4 A simplified diagram illustrating the three-position process of the shape adjustment mechanism in the embodiment;

[0043] Figure 5 This is a schematic diagram illustrating the dynamic center of gravity compensation mechanism within the outer layer of the implementation method.

[0044] 1-Robot chassis; 2-Main working nozzle; 3-Negative pressure induction nozzle; 4-Lateral adjustment nozzle; 5-Support frame; 6-Adjustment motor; 7-Eccentric rounded rectangular cam; 8-Wear-resistant rubber sleeve; 8a-Flexible covering layer; 9-Longitudinal slide rail; 10-Battery pack; 11-Reset spring; 100-Inner layer; 200-Outer layer. Detailed Implementation

[0045] Example 1:

[0046] Combination Figure 1 A de-icing jet control method for a rotary blade de-icing robot, wherein the robot is equipped with a fluid distribution component, the input end of which is connected to a high-pressure fluid source, and its output end is at least branched into a first jet end facing the de-icing operation path and a second jet end facing the converging curved surface at the bottom of the robot body; the control method includes:

[0047] The system acquires the combined external load vector generated by the robot under the rotating reference frame in real time. The combined external load vector is at least composed of the robot's gravity, environmental aerodynamic forces, and rotational inertial forces coupled together. Based on the parameters of the de-icing jet to be executed, the system estimates the transient recoil vector generated at the moment of triggering the first jet. According to the spatial vector relationship between the combined external load vector and the transient recoil vector, the system calculates the jet phase window where the robot's adhesion torque margin meets a preset threshold. If it is determined that the robot is currently in the jet phase window and the de-icing operation is triggered, the system controls the fluid distribution component to perform transient multi-channel fluid distribution: directing a first proportion of high-pressure fluid to the first jet to output an ice-breaking jet, and simultaneously directing a second proportion of high-pressure fluid to the second jet to drive the high-pressure fluid to sweep tangentially along the convergent surface at the bottom of the robot to generate transient aerodynamic downforce pointing towards the working surface.

[0048] When implementing this, the following steps can be followed:

[0049] Step 101: Obtain robot operating parameters in real time; obtain the robot body attitude angle through the inertial measurement unit, and obtain the fan rotor angular velocity and robot operating radius.

[0050] Step 102: Obtain preset de-icing operation parameters; read the currently set physical parameters such as the main jet target pressure and nozzle cross-sectional area.

[0051] Step 201: Synthesize the comprehensive external load vector; superimpose the gravity vector, centrifugal force vector, and environmental aerodynamic force vector in the spatial dimension.

[0052] Step 202: Estimate the transient recoil vector; perform forward calculation based on the target pressure of the main jet to calculate the magnitude and direction of the recoil force generated at the moment of triggering at the first jet end.

[0053] Step 301: Calculate the fuselage adhesion moment margin; determine the potential overturning axis of the fuselage, and calculate the residual constraint moment generated by the combined external load and recoil vector on this axis.

[0054] Step 302: Determine if the system is within the jet phase window; determine if the current torque margin is greater than the preset safety reserve threshold; if "No": return to Step 101, the system continues to monitor and wait. If "Yes": proceed to the next step and trigger the jetting operation.

[0055] Step 401: Calculate the dynamic allocation command; calculate the target aerodynamic pressure in reverse based on the torque balance constraint, and determine the allocation values ​​of the first ratio and the second ratio.

[0056] Step 402: Perform transient multi-channel fluid distribution; drive the fluid distribution component to operate, and synchronously guide the high-pressure fluid to different injection ends in proportion.

[0057] Step 403: Activate pneumatic downforce; the first jet outputs an ice-breaking jet, and the fluid from the second jet simultaneously sweeps tangentially along the converging surface of the abdomen, generating downforce pointing towards the working surface.

[0058] Step 501: Real-time monitoring of body displacement and adhesion pressure; using a body height sensor to monitor displacement deviation and simultaneously obtain the actual adhesion pressure of each suction cup foot.

[0059] Step 502: Extract loss features and reconstruct the model; extract compensation deviations based on the fuselage's minute displacement or the decrease in adhesion, generate adaptive correction factors, and reconstruct the mapping model.

[0060] Step 503: Dynamically adjust the allocation ratio; adjust the second ratio in real time according to the reconstruction model to achieve closed-loop compensation of downforce.

[0061] Step 504: End the current blowing cycle; complete this operation and the system returns to the starting step to enter the next cycle.

[0062] In implementing this solution, the de-icing robot senses the environment through its built-in IMU and distance encoder, calculating the rotation phase and overall external load. The system predicts the recoil vector and substitutes it into the dynamic stability model to find the jet phase window. During the jet execution phase, a high-speed reversing valve splits the high-pressure airflow into two paths: one for de-icing and the other directed to the underside nozzle. The underside curved surface and the blade surface form a flow channel, generating aerodynamic downforce according to Bernoulli's principle, balancing the recoil torque, achieving self-balancing, and improving adhesion stability.

[0063] In implementing the control process of this scheme, the central processing unit first establishes a three-dimensional dynamic model of the robot through a multi-sensor fusion algorithm. The system acquires the fuselage attitude angle sequence from the six-axis inertial measurement unit (IMU) through high-frequency sampling (the sampling frequency is usually not less than 100Hz), including pitch angle θ, roll angle ϕ, and yaw angle ψ, and uses a coordinate transformation matrix to project the gravitational acceleration constant g onto the fuselage coordinate system to obtain the gravity vector. The components of the three axes of the machine body. Simultaneously, the system reads the rotor angular velocity ω of the wind turbine main control system in real time via industrial Ethernet (such as EtherCAT or CANopen), and, in conjunction with a laser rangefinder or encoder installed on the bottom of the robot, obtains the radial distance r of the robot relative to the center of the hub, and then uses the formula... Calculate the centrifugal force vector. This refers to a vector pointing from the center of rotation (i.e., the center point of the wind turbine hub) to the robot's center of mass (or geometric center), with a magnitude equal to the straight-line distance between the robot and the center of the hub. This is combined with the environmental aerodynamic vector calculated from the chassis pressure sensors. The final synthesized external load vector of the robot in the rotating reference frame + , The centrifugal force vector is calculated by taking the robot's radial coordinate r on the blade, obtained from the hub distance encoder or heading calculation, the robot's mass m, and the current angular velocity ω of the wind turbine, to determine the magnitude of the centrifugal force (its vector direction is from the center of the hub to the tip of the blade). .

[0064] In the pre-judgment phase before pulse intervention, the system executes a forward calculation process: the central processing unit retrieves the target pressure P of the main jet set for the current task. s and the cross-sectional area A of the nozzle at the first injection end s The transient recoil vector generated at the moment of triggering is calculated using the momentum mapping formula. To determine fuselage stability, the system first identifies the potential overturning axis L formed by the support lines of the rear fuselage attachment units, and uses this as a reference to calculate the combined external load moment. The system searches for the torque balance criterion by traversing the load evolution trend across the entire 360-degree phase domain of rotor rotation. The phase interval is defined and marked as the jet phase window. To provide a reference adhesion torque for the vacuum suction cup foot, This represents the recoil and overturning moment generated by the transient recoil vector, which causes the fuselage to pitch up. This indicates the preset lower limit of the safety reserve torque to ensure stable adhesion. This represents the momentum mapping coefficient, used to characterize the proportional relationship between fluid pressure and mechanical recoil force.

[0065] When the rotor phase enters the preset window and triggers a de-icing command, the system switches to a reverse calculation process to determine the compensation parameters. The central processing unit first calculates the target aerodynamic downforce required to neutralize the nose-up moment based on the balance constraints. Subsequently, the system invokes a preset hydrodynamic acceleration mapping model of the omnidirectional convergent surface at the bottom of the fuselage ( Figure 2 The direction of the middle arrow indicates the direction of the external airflow. By using an omnidirectional convergent surface (which can be spherical) to create different flow velocities above and below the robot, this model describes the relationship between the pressure difference and flow velocity in the flow channel based on Bernoulli's principle. Using this model, the processor inversely calculates the tangential fluid target velocity required to generate the target negative pressure within the convergent surface. This allows for the calculation of the required fluid flow rate at the second injection end. This represents the aerodynamic pressure difference formed inside and outside the convergent surface (negative pressure generated by the Venturi effect). This indicates the real-time air density in the working environment. This indicates the tangential fluid target velocity of the fluid generated at the second injection end within the converging curved flow channel. This indicates the relative wind speed outside the fuselage relative to the working surface.

[0066] The central processing unit calculates the dynamic allocation instructions (such as the first proportional valve) for the multi-way proportional directional valve group in the fluid distribution component based on the flow requirements of the first and second injection ends. Compared to the second ratio (Duty cycle signal). At the instant the main jet is ejected, the negative pressure induction nozzle's jet path obliquely sweeps across the fuselage belly and tangentially induces the negative pressure field at the geometric extreme point, forming a high-speed flow field in the belly channel. In specific implementation, the system can perform correlation analysis by monitoring fuselage displacement and adhesion pressure loss to extract the downpressure compensation deviation; as an optional solution, the aerodynamic pressure distribution field can also be obtained through a pressure sensing matrix distributed within the curved surface. If the measured value deviates from... The system reconstructs the gain coefficient of the mapping model through an adaptive correction factor α, dynamically adjusting the opening of the proportional valve. This mathematical processing flow based on spatial vector relationships ensures that the transient aerodynamic downpressure can be precisely boosted according to the intensity of the ice-breaking jet, thus mitigating the tendency of the fuselage to pitch up and overturn.

[0067] Example 2

[0068] Based on the above embodiments, this solution calculates the jet phase window where the fuselage adhesion torque margin meets a preset threshold according to the spatial vector relationship between the comprehensive external load vector and the transient recoil vector. The steps include: determining the potential overturning axis when the transient recoil vector acts on the fuselage, causing the fuselage to tend to tilt relative to the working surface; using the potential overturning axis as a reference, calculating the comprehensive external load torque sequence dynamically mapped by the comprehensive external load vector within the 360-degree full-phase domain of the fan rotor rotation, and calculating the recoil overturning torque generated by the transient recoil vector relative to the potential overturning axis; traversing the comprehensive external load torque sequence and selecting the phase intervals that make the comprehensive external load torque manifest as a stable clamping torque pointing towards the working surface; within the phase interval, if the sum of the stable clamping torque and the effective adsorption torque provided by the current robot's attachment unit, minus the recoil overturning torque, results in a remaining constraint torque greater than a preset safety reserve threshold, then the corresponding continuous phase interval is marked as the jet phase window.

[0069] In practice, the central processing unit executes a pre-control algorithm based on time series prediction, which aims to find a safe trigger window for the lightweight robot to blow air in a variable-direction composite force field. Specifically, it includes the following operations.

[0070] The system needs to clearly define the physical point of application of the load interference, that is, the central processing unit combined with the current geometry of the fuselage and the real-time estimated transient recoil vector. The potential overturning axis L for the fuselage to tilt up or backward is determined. In actual operation, when the main working nozzle blows forward, axis L is usually defined as the line connecting the center positions of all vacuum suction cup foot support points in the adsorption state at the reverse end (rear end) of the fuselage. The next step is to enter the full-phase torque simulation stage; the central processing unit establishes a three-dimensional torque balance model with the potential overturning axis L as the coordinate origin through the built-in dynamics engine. The system uses an absolute encoder installed at the wind turbine hub to obtain the real-time rotation angle Φ∈[0°,360°] of the wind turbine rotor, and uses this as a variable to calculate the comprehensive external load vector. The dynamic mapping process occurs within a complete rotation cycle. Specifically, the system generates a comprehensive external load torque sequence. (Φ), its calculation formula is:

[0071] + )× .

[0072] in, This is the lever arm vector of the point of application of the resultant force of the combined external loads relative to the potential overturning axis L. Simultaneously, the system adjusts the lever arm vector based on the estimated transient recoil vector. Calculate the resulting recoil and overturning moment. :

[0073] = ×

[0074] in, This is the vertical height vector of the jet center axis at the first injection end relative to axis L. Next, the system performs logical traversal and filtering. The central processing unit traverses and generates... The sequence identifies all phase intervals where the torque direction points towards the working surface. Within these intervals, the resultant force of natural loads (gravity and centrifugal force) can exert a clamping effect on the fuselage; the corresponding torque is defined as the stable clamping torque. The system then applies the effective adsorption torque provided by the attachment unit. Introduce a decision equation.

[0075] Obtained through the following equation:

[0076] =

[0077] in, Let S be the real-time vacuum pressure of the i-th suction cup in the adsorption state, and S be the suction cup area. Let L be the lever arm distance from the geometric center of the suction cup to the axis L. Then, the system determines the final jet phase window by performing a safety comparison of the remaining constraint torques. Within the selected stable clamping torque phase range, the processor verifies the following inequality:

[0078]

[0079] Under a specific phase, the combined external load manifests as a stable clamping torque pressing against the working surface; It is a torque component generated by the synthesis of gravity vector, rotational inertia force vector and environmental aerodynamic force vector when the wind turbine rotates to a specific angle (phase Φ), which is spatially directed towards the working surface and can suppress the tendency of the turbine to detach. This is a preset safety reserve torque value used to ensure attachment redundancy; if this inequality holds, it means that when the pulse jet is triggered at this phase point, even if subjected to a high-energy recoil impact, the fuselage still has sufficient torque margin to counteract the nose-up tendency. The central processing unit marks the continuous phase interval that meets this condition as the jet phase window. In practical operation, the system dynamically adjusts based on feedback from wind speed and ice thickness. If the current state is in a step displacement state, the number of suction cups n involved in the calculation will decrease accordingly, resulting in As the pressure decreases, the system automatically contracts or suspends the jet phase window. Through this full-phase mapping processing based on the potential overturning axis, this scheme can identify the optimal phase during the wind turbine's rotation when gravity and rotational inertia are in a natural compressive state. This proactively avoids the risk of overturning caused by the superposition of impact loads and natural loads in the time dimension, providing solid algorithmic support for the stable online operation of de-icing robots, especially lightweight de-icing robots with base stations.

[0080] The steps of controlling the fluid distribution component to perform transient multi-channel fluid distribution and dynamically adjust the first and second ratios include:

[0081] The system obtains the target pressure of the main jet required for the current de-icing operation, and calculates the transient recoil torque generated at the moment of triggering based on the target pressure of the main jet and the nozzle cross-sectional area of ​​the first jet end. Based on the torque balance constraint, it calculates the target aerodynamic downforce required to counteract the transient recoil torque. It calls the preset fluid dynamic acceleration mapping model of the convergent surface at the bottom of the fuselage to calculate the tangential fluid target velocity required to generate the target aerodynamic downforce within the convergent surface. According to the fluid consumption corresponding to the target pressure of the main jet and the fluid consumption corresponding to the tangential fluid target velocity, it calculates the dynamic allocation commands of the first ratio and the second ratio, and outputs them to the drive control terminal of the fluid allocation component. Among them, the greater the target pressure of the main jet, the greater the fluid kinetic energy of the second ratio allocated to the second jet end, which is used to make the transient aerodynamic downforce generated at the bottom of the fuselage increase with the intensity of the ice-breaking jet at the first jet end.

[0082] In the specific implementation of de-icing jet control, the system first executes a forward calculation process based on the physical input. Within the period preceding each pulse command trigger, the central processing unit acquires the target pressure P of the main jet set for the current de-icing task. s Combined with the pre-stored nozzle cross-sectional area A of the first injection end (main nozzle). s The system constructs a transient recoil force calculation model:

[0083]

[0084] Where λ is the momentum transfer correction coefficient. The system incorporates the geometric lever arm L of the nozzle relative to the fuselage's center of mass.arm The transient recoil torque generated at the moment of triggering is derived. This allows the system to pre-calculate the magnitude of the disturbance that will impair the fuselage's adhesion stability, given the known jet intensity as input. The system then enters the countermeasure solution phase. Based on the torque balance constraint, the central processing unit calculates the parameters necessary for substantial neutralization. Required target aerodynamic downforce The system then switches to the reverse calculation process, calling the preset hydrodynamic acceleration mapping model of the omnidirectional convergent surface at the bottom of the fuselage. This model, based on Bernoulli's principle, describes the nonlinear mapping relationship between the pressure drop (negative pressure) and the flow velocity in the flow channel:

[0085]

[0086] Where S is the effective projected area of ​​the convergent surface. The central processing unit will... Substituting the objective value into the model for inverse solving, the tangential fluid target velocity required to generate suction of this magnitude within the convergent curved flow channel is calculated. In order to achieve reverse reasoning.

[0087] Ultimately, the central processing unit (CPU) determines the final decision based on P. s The corresponding main jet fluid consumption Q1 and the amount of fluid consumed by the main jet are related to the fluid consumption of the main jet. Based on the calculated secondary jet fluid consumption Q2, the drive control command for the fluid distribution component is calculated. This command is output to the multi-way proportional directional valve group in the form of a first proportional n1 and a second proportional n2:

[0088] n1:n2=Γ(Q1,Q2)

[0089] Where Γ is the valve flow characteristic mapping function.

[0090] In terms of implementation effectiveness, this control logic exhibits a follow-up pressurization characteristic: when the required pressure for de-icing is reached... s When the force increases, the system automatically identifies the recoil torque M. r The synchronous enhancement, and then through algorithmic linkage, forcibly increases the second proportional fluid kinetic energy allocated to the second jet end, thereby increasing the transient aerodynamic downpressure generated in the abdomen. This constructs a self-soothing system that counteracts the head-up tendency caused by the jet recoil, ensuring that the lightweight robot can maintain a stable force posture under impact from pulses of different energy levels.

[0091] This embodiment uses forward calculation to predict disturbances based on input conditions, and backward calculation to solve for countermeasure parameters based on the target. The solution combines both methods to achieve parameter matching with requirements. The higher the pressure of the main nozzle, the stronger the underbody pressure, forming a self-stable system to ensure fuselage stability and reduce the risk of rollover.

[0092] In implementing the adaptive closed-loop correction steps of this invention, the system monitors the final result of the robot's attachment state, reversely evaluates and corrects the pneumatic boosting effect, thereby avoiding the reliability problem of directly deploying sensors in harsh abdominal environments.

[0093] The system uses laser rangefinders located at the four corners of the fuselage to acquire the displacement deviation h of the fuselage relative to the working surface in real time. off Simultaneously, the central processing unit acquires the negative pressure value and rate of change of each foot end through pressure sensors integrated in the negative pressure pipelines of each attachment unit (vacuum suction cup foot), as attachment status parameters. During the de-icing pulse triggering period, if the body experiences a slight upward or deflection tendency due to insufficient pneumatic pressure, the laser rangefinder will detect the instantaneous increase in the value. At the same time, a slight air leakage inside the suction cup will cause fluctuations in the negative pressure value.

[0094] The process involves extracting loss features and correlating them with the deviation. The central processing unit (CPU) then processes the acquired displacement deviation value h. off The current adhesion performance loss characteristics are extracted by temporal coupling with the adhesion pressure fluctuation. The processor then correlates these characteristics with a preset target aerodynamic downpressure F. tg Perform correlation analysis (the central processing unit will determine the current target aerodynamic downforce F) tg (As a benchmark reference value, and correlated with the measured adhesion state parameters): If it is determined that the fuselage has experienced a displacement increment, it proves that the actual downforce generated has failed to effectively reduce the transient recoil vector. The system then calculates the downforce compensation deviation value accordingly.

[0095]

[0096] Where Ψ is the proportional mapping function, used to characterize the linear or nonlinear weighted contribution of displacement deviation and pressure loss relative to the target pressure. The physical meaning of this formula is: the system determines the original... What percentage of effectiveness is missing? This represents the vacuum loss value inside the suction cup.

[0097] Next, the system performs model reconstruction. The central processing unit substitutes the deviation value ϵ into the adaptive control operator to generate an adaptive correction factor α used to correct the deviation of the physical model. Specifically, the formula for calculating the adaptive correction factor is: Where 1 is a reference constant, This is the proportional gain coefficient. The integral gain coefficient is denoted as dt, where dt is the time derivative term. This factor directly affects the pre-defined hydrodynamic acceleration mapping model; by adjusting the gain coefficient C in the model... k Perform real-time reconstruction, i.e., update to .

[0098] By reconstructing the gain coefficient, the system actually compensates for energy conversion losses caused by uneven working surfaces, air leaks, or fuselage attitude deflection at the algorithm level, enabling the physical model to self-evolve based on the real-time feedback of the attachment results.

[0099] Finally, the central processing unit inputs the reconstructed mapping model, recalculates, and dynamically updates the second proportional command of the fluid distribution component. By fine-tuning the opening of the proportional reversing valve assembly, the system automatically increases the kinetic energy of the fluid diverted to the second injection end to enhance the flow velocity within the omnidirectional convergent surface of the abdomen, until the body displacement deviation returns to zero and the suction cup negative pressure stabilizes. This closed-loop correction mechanism based on result feedback ensures that the robot can maintain a highly reliable attachment state through adaptive follow-up pressurization even when facing non-ideal contact surfaces or complex directional disturbances.

[0100] Through a closed-loop correction mechanism that uses fuselage displacement deviation and adhesion pressure feedback, the system can detect aerodynamic pressure fluctuations caused by uneven surface roughness, residual ice interference, or fuselage attitude deflection in real time, and dynamically compensate for the flow split ratio with an adaptive correction factor. This closed-loop feedback control improves system robustness, ensuring that the expected targeted aerodynamic downforce can still be generated on the fuselage bottom even on non-ideal contact surfaces, thus achieving high reliability in maintaining the adhesion state.

[0101] Furthermore, the control method also includes dynamic compensation steps for the atmospheric physical environment, including:

[0102] The system acquires atmospheric pressure and ambient temperature in real time, calculates air density correction coefficients under current altitude and meteorological conditions, calibrates parameters of the fluid dynamics acceleration mapping model using the air density correction coefficients, and dynamically corrects the mapping function between tangential fluid target velocity and transient aerodynamic pressure according to changes in the atmospheric physical environment. The second scale is updated based on the corrected mapping function, thus enabling precise control of transient aerodynamic pressure on transient recoil vector under different atmospheric density conditions.

[0103] During task initialization or climbing operations, the central processing unit acquires the atmospheric pressure P of the working environment in real time through an environmental sensing module (including a digital barometer and a high-precision thermistor) integrated into the robot body or supply base station. apa The ambient temperature T is considered. Given the wide geographical distribution of wind farms, ranging from high-altitude mountain wind farms to near-shore offshore wind farms, the atmospheric environment varies. The air density correction factor ρ is calculated for the current altitude and meteorological conditions. corr The calculation logic is as follows:

[0104]

[0105] in, and These are standard atmospheric pressure and standard temperature constant, respectively. P represents the real-time air density under current operating conditions. a T represents the real-time atmospheric pressure of the operating environment as collected by the sensor. T represents the real-time temperature in Celsius of the operating environment as collected by the sensor. The air density correction factor characterizes the current environment relative to standard operating conditions. The preset standard atmospheric pressure (e.g., 101.325 kPa). Standard temperature (e.g., 288.15K).

[0106] Subsequently, the system enters the parameter calibration process. The central processing unit will calculate ρ... corr Substituting this into the aforementioned fluid dynamics acceleration mapping model, we find that the transient aerodynamic downpressure generated by the Venturi effect is linearly positively correlated with air density, i.e., satisfying... ∝ρ⋅ The physical laws governing the robot's operation in a high-altitude, low-pressure environment ( If the original tangential fluid target velocity is maintained, the resulting downforce will be insufficient to offset the transient recoil vector. Therefore, the system dynamically compensates for the mapping function by adjusting the correction coefficients and recalibrates to generate the tangential fluid target velocity. With transient aerodynamic pressure The numerical correspondence between them.

[0107] In practice, this process manifests as automatic optimization of the control strategy: in high-altitude areas with thin air, the system automatically increases the proportion of fluid kinetic energy diverted to the second jet end, compensating for pressure loss due to density reduction by increasing the tangential velocity of the airflow within the channel; while in offshore wind fields with higher air density, the proportional allocation is optimized accordingly to avoid energy redundancy caused by overcompensation. Through this dynamic calibration based on the atmospheric physical environment, this solution ensures that the robot can achieve accurate balancing of transient aerodynamic downpressure and transient recoil vector under different latitudes, altitudes, and seasonal conditions worldwide.

[0108] Since the aerodynamic downforce generated by the Venturi effect is highly dependent on air density, this solution can automatically adjust the fluid distribution strategy according to altitude (such as mountain wind fields and offshore wind fields) and climate change, reducing the possibility of undercompensation or overcompensation caused by changes in the atmospheric environment. This technology ensures that the robot can maintain a constant and accurate load cancellation capability at different latitudes, longitudes and altitudes around the world, greatly expanding the operational boundaries of the system.

[0109] Example 3

[0110] Based on the above embodiments, the output end of the fluid distribution component further branches into third injection ends pointing to the attachment units on the left and right sides of the fuselage respectively; the control method also includes a lateral correction step for smoothing yaw slip, including:

[0111] The rotational angular velocity of the fan rotor and the displacement velocity of the robot are acquired in real time. The Coriolis force component generated by their coupling is calculated, and combined with the lateral eccentricity of the transient recoil vector, a transient yaw disturbance torque is synthesized. According to the spatial deflection direction and amplitude of the transient yaw disturbance torque, the fluid distribution component is controlled to perform asymmetric differential fluid distribution to the third jet end located on the left and right sides of the fuselage. The third jet end on the same side as the deflection trend of the transient yaw disturbance torque is controlled to instantaneously increase the flow split ratio and output a lateral pressurized jet. The recoil momentum generated by the lateral pressurized jet is used to construct a lateral correction torque that dynamically counteracts the transient yaw disturbance torque. Simultaneously, the lateral pressurized jet is used to blow away the derived water film and residual ice between the attachment unit on this side and the working surface to improve the anti-slip static friction boundary of the attachment interface.

[0112] During implementation, the central processing unit establishes a lateral force model through multi-source data fusion. The system obtains the rotor rotational angular velocity ω in real time from the wind turbine main control unit and, combined with the real-time travel speed v fed back by the robot encoder, calculates the Coriolis force component F caused by the rotating reference frame. cor Its calculation formula is: F cor =2⋅m⋅ω⋅v⋅sin(γ). Where γ is the angle between the robot's travel direction vector and the rotor's rotation axis vector, and m is the robot's mass. Simultaneously, the system analyzes the spatial orientation of the transient recoil vector. If the jet centerline of the first injection end (main nozzle) has a lateral eccentricity distance e relative to the geometric centerline of the fuselage, an additional recoil eccentric force will be generated. The central processing unit vector-superimposes the torque generated by the Coriolis force component with the torque generated by this eccentric load to synthesize the final transient yaw disturbance torque M. yaw This torque reflects the physical tendency of the robot to rotate or deviate from its predetermined trajectory around a central axis perpendicular to the working surface at the instant of pulse triggering.

[0113] Subsequently, the system executes an allocation strategy based on asymmetric logic. Once the central processing unit detects the transient yaw disturbance torque M... yaw The direction and amplitude are immediately given to the fluid distribution component, which then issues an instruction to implement asymmetrical differential fluid distribution to the third injection end (lateral auxiliary nozzle) located on the left and right sides of the fuselage.

[0114] The fluid distribution component controls the fluid pressure and flow rate distributed to the left and right third injection ends by adjusting the independent output branches of its internal multi-way proportional directional valve. Taking a leftward yaw tendency of the fuselage as an example, the processor instructs the third injection end on the same side as the yaw tendency (i.e., the left side) to instantaneously increase the flow split ratio. At this time, the lateral pressurized jet ejected from the nozzle on this side will generate a reverse lateral thrust on the fuselage due to the principle of conservation of momentum. Since this lateral thrust acts on the edge of the fuselage, it can construct a lateral correction torque that is opposite in direction and equal in magnitude to the transient yaw disturbance torque, thereby reducing the yaw tendency of the fuselage to zero within milliseconds. During the synchronous execution of the correction action, this solution also achieves friction enhancement compensation at the physical interface. Through the specialized installation angle of the third injection end, the high-speed jet path is made to obliquely sweep across the gap between the bottom of the corresponding attachment unit (vacuum suction cup foot) and the working surface. The airflow can instantly sweep away and peel off the derived water film, residual ice and debris in the area, reducing the lubrication effect between interfaces on a microscopic level and significantly improving the anti-slip static friction boundary of the attachment interface.

[0115] Example 4

[0116] Based on the above embodiments, the control method further includes a coordinated interlocking step between the jet and the gait, specifically comprising:

[0117] The robot monitors the operating phase of each attachment unit in real time to determine whether the robot body is currently in a stepping phase where some attachment units are detached from the working surface, or in a supporting phase where all attachment units are attached to the working surface. An interlocking mechanism is established between the jet trigger command for de-icing operations and the operating phase: if the robot body is determined to be in a stepping phase, the jet trigger command is suspended, restricting the fluid distribution component from performing transient multi-channel fluid distribution until the robot body returns to the supporting phase; if the robot body is determined to be in the supporting phase and de-icing operations are triggered, during the synchronization of the fluid distribution component performing transient multi-channel fluid distribution, the drive joints of each attachment unit are controlled to enter a rigid locking mode; a rigid transmission path is constructed using the attachment units in rigid locking mode to penetrate the robot body and release the impact load caused by the transient recoil vector to the working surface.

[0118] During implementation, the central processing unit (CPU) utilizes pressure sensors integrated at the ends of each attachment unit and high-precision position encoders on the drive joints to construct a real-time operational phase characteristic map of the robot. The CPU uses a preset logical judgment algorithm to poll the ground-adherence status of each foot at millisecond intervals: if the real-time attachment pressure of any set of attachment units is detected to be below a safety threshold, or if its drive joint is executing a position change command, the system determines that the robot is currently in the stepping phase; if all attachment unit pressure sensors report stable positive pressure, and the encoder readings of each joint remain constant, the system determines that the robot is currently in the support phase.

[0119] Subsequently, the central processing unit (CPU) establishes a mandatory constraint relationship between the jet trigger command and the operating phase through a software-defined command gate. In actual operation, even if the system calculates a relatively optimal jet phase window based on the aforementioned algorithm, if the gait logic feedback indicates a stepping phase (i.e., part of the suction cup foot detaches from the working surface), the system will immediately activate the command suspension protocol. At this time, the drive signal of the fluid distribution component is physically shielded or logically blocked, restricting fluid output from all jet ends until the gait control sequence completes the stepping cycle and enters the support phase with all feet grounded. This interlocking mechanism ensures that the robot is always in a physical state of maximum adhesion stiffness when performing high-energy pulse actions.

[0120] During the synchronization process of confirming entry into the support phase and issuing the emission trigger command, the central processing unit (CPU) sends a rigidity enhancement command to the drive joints of each attached unit, driving each active drive joint to instantly enter a rigid lock-up mode. If the drive joint uses an electric servo system, the CPU instantly adjusts the position loop gain in the motor's PID control parameters to its maximum value, or physically locks the motor shaft through an electromagnetic brake mechanism. If the drive joint uses a hydraulic mechanism, a hydraulic lock is formed by shutting off the control valve of the hydraulic circuit, transforming the robot from a flexible body with obstacle avoidance flexibility into a near-rigid structure.

[0121] Finally, by engaging the attachment unit in rigid locking mode, the system constructs a rigid transmission path within the fuselage, extending from the first injection end (main nozzle) through the frame, locking joint, and down to the bottom of the suction cup foot. When the transient recoil vector generated by the de-icing pulse... Upon impact with the fuselage, the impact load no longer causes slight elastic deformation or positional displacement of the mechanical joints. Instead, it penetrates directly through the fuselage along this transmission path and is released onto the massive wind turbine blade substrate. This effectively suppresses low-frequency resonance and fuselage flutter caused by the flexible gaps in the joints, ensuring that the fuselage remains stationary at the moment the ice-breaking pulse is released. This achieves synergy between the high-energy de-icing action and the climbing stepping action in terms of timing, stiffness, and safety. The aforementioned stepping phase refers to the state where some attached units are detached from the working surface and swing forward during the robot's displacement cycle. The support phase refers to the statically stable state where all attached units of the robot are sealed and adhered to the working surface. The rigid transmission path refers to the load transfer channel without elastic displacement formed between the force points of the fuselage and the working surface by locking the joint stiffness.

[0122] Example 5

[0123] Based on the above embodiments, the control method further includes a dynamic protection step based on the coupling of thermal load and aerodynamic pressure, including: real-time acquisition of the ambient temperature and the status of the fuselage system, the system status including at least the drive joint resistance and edge sealing margin of each attachment unit; if it is determined that the ambient temperature is lower than a preset temperature threshold, the fluid distribution component is controlled to guide the thermo-pressure fluid generated by the high-pressure fluid source to perform anti-icing heat maintenance operation: a set proportion of thermo-pressure fluid is directed to the edge guide gaps of the drive joint and attachment unit; if it is determined that the comprehensive external load vector touches the instability boundary during the de-icing operation, an energy redistribution action is performed: a forced switch to the second distribution ratio is performed, and the full amount of thermo-pressure fluid is directed to the second injection end.

[0124] The system assesses environmental threats and the health of the machine body through a real-time monitoring mechanism. The central processing unit acquires the operating ambient temperature T collected by the environmental sensing module in real time and simultaneously polls the active drive joint current feedback value of each attachment unit. Since the viscosity of the lubricating grease increases at low temperatures, which can cause an abnormal increase in motor current, the system calculates the drive joint resistance accordingly. At the same time, the processor monitors the vacuum fluctuation rate at the edge of each suction cup through pressure sensors in each attachment unit to assess the edge sealing margin.

[0125] During the normal anti-freeze implementation phase, if the temperature T is determined to be below a preset temperature threshold (e.g., zero degrees or below), the central processing unit controls the fluid distribution component to enter the heat maintenance mode. At this time, the system guides the thermocompressed fluid (whose temperature is typically between 50°C and 80°C) generated by the high-pressure fluid source during compression for directional flow. The system controls the proportional reversing valve assembly to deliver a set proportion of thermocompressed fluid through the insulated micro-tubes embedded in the body to the sealed cavities of each drive joint, using the thermal conduction effect to suppress the freezing of the lubricating medium and ensure the responsiveness of gait execution. Simultaneously, another portion of the thermocompressed fluid is guided to the edge guide gaps at the edge of the attachment unit, where the ejected hot airflow forms a transient thermal convection layer at the suction cup sealing interface, thereby reducing the risk of air leakage caused by condensation and frost due to temperature differences and maintaining the high sealing performance of the attachment interface.

[0126] During the critical pressurization execution phase, if the robot is performing pulse jet injection, the central processing unit determines the current comprehensive external load vector. When the rate of change of the modulus or the pointing angle reaches the preset instability boundary (i.e., when the remaining torque margin approaches the safety reserve threshold), the system will trigger the highest priority energy redistribution action. At this time, the central processing unit issues an instruction to forcibly switch the fluid distribution ratio, melting all the first proportion of fluid originally used for front-end de-icing and the portion of fluid used for heat maintenance, and instead directing the full amount of thermopressurized fluid to the second jet end located at the geometric extreme point of the abdomen.

[0127] The injection of the full volume of fluid causes the tangential flow velocity within the omnidirectional convergent surface of the abdomen to instantly jump to the design limit, according to the aforementioned fluid dynamics acceleration mapping model:

[0128]

[0129] v is the maximum transient aerodynamic downforce that the system can generate. max The limiting tangential flow velocity within the abdominal flow channel under conditions of full energy distribution.

[0130] The system utilizes the real-time air density that reaches its peak at this time. With limiting flow velocity v max This generates the maximum transient aerodynamic downpressure. This energy redistribution mechanism sacrifices instantaneous de-icing progress and heat maintenance to achieve highly reliable adhesion constraints on the fuselage under extreme sudden loads (such as instantaneous strong gusts or pulse recoil), thus achieving emergency protection. Through this deep coupling and control of thermal and pressure energy fields, this solution enhances the robot's continuous operation capability in extremely cold environments and its physical resistance boundary to sudden risks. The aforementioned edge guide slots are microchannels located at the edge of the suction cup used to expel hot airflow and suppress frost formation.

[0131] Example 6

[0132] Combination Figures 2 to 5 A de-icing spraying system for a rotary blade de-icing robot is disclosed. The system includes: a fluid control hub, a robot chassis 1, a nozzle array, a main working nozzle 2, a negative pressure induction nozzle 3, a lateral adjustment nozzle 4, and a central processing unit. The fluid control hub includes a high-pressure fluid inlet and a multi-way proportional directional valve assembly connected to the high-pressure fluid inlet. The underside of the robot chassis 1 has an omnidirectional convergent surface, which smoothly bulges inward from the peripheral edge of the chassis 1, and the geometric extreme point of the omnidirectional convergent surface is located in the central region of the underside of the chassis. The nozzle array is distributed at different phases of the chassis, including the main working nozzle 2, the negative pressure induction nozzle 3, and the lateral adjustment nozzle 4. The main working nozzle 2 is located at the front of the machine body and points to the working path through the first output end of the multi-way proportional directional valve group; the negative pressure induction nozzle 3 is configured to point tangentially at the geometric extreme point of the omnidirectional convergent surface and is supplied with air by the second output end of the multi-way proportional directional valve group, and the spray axis of the negative pressure induction nozzle 3 is parallel to the tangent of the omnidirectional convergent surface at the geometric extreme point; the lateral adjustment nozzles 4 are symmetrically arranged on the left and right sides of the machine body and point to the attachment units on both sides of the machine body, and the two lateral adjustment nozzles 4 are supplied with air through the two independent third output ends of the multi-way proportional directional valve group; the central processing unit is electrically connected to the multi-way proportional directional valve group and is used to execute the above-mentioned de-icing spray control method.

[0133] Specifically, the robot body corresponding to this de-icing and blowing system can be integrally molded from lightweight, high-strength composite materials (aluminum alloy or fiberglass). The belly of the robot chassis 1 is constructed as a specific omnidirectional converging surface. Geometrically, this surface presents an inverted shallow bowl-shaped or spherical crown-shaped structure that gradually and smoothly bulges inward from the peripheral edge of the chassis. Its geometric extreme point (i.e., the highest vertex of the bulge) is located in the central region of the belly of the chassis, ensuring that the resultant point of the generated downward pressure coincides with the geometric center of the chassis. The edges of the chassis can have smooth chamfered transitions to guide the external natural wind field or rotationally induced wind smoothly into the belly flow channel.

[0134] The core component of this system is an integrated high-pressure pneumatic valve block, which can be equipped with a high-pressure fluid inlet for connecting the umbilical cable and is internally connected to a multi-way proportional directional valve assembly. This valve assembly contains multiple high-speed proportional electromagnetically driven valve cores, capable of independently adjusting the flow distribution of multiple output terminals within milliseconds based on the pulse width modulation (PWM) signal from the central processing unit. The valve block leads out a first output terminal (connected to the main working nozzle 2), a second output terminal (connected to the negative pressure induction nozzle 3), and two independent third output terminals (connected to the two lateral adjustment nozzles 4) via internal manifolds.

[0135] The main working nozzle 2 is located at the front end of the machine body in the direction of travel. Its spray port is slightly tilted downwards and points towards the surface of the blade ice layer to be treated. The highest energy level fluid is obtained through the first output end of the multi-way proportional reversing valve group to perform stripping and crushing operations.

[0136] The three negative pressure induction nozzles are arranged at the front of the fuselage, and their jet path obliquely sweeps across the belly of the fuselage and tangentially induces the negative pressure field at the geometric extremum point. During operation, the high-pressure fluid is ejected radially from the extremum point along the curved surface, and the Coanda effect is used to make the airflow flow at high speed close to the converging curved surface.

[0137] Lateral adjustment nozzles 4 are symmetrically mounted on the side edges of the left and right wings of the fuselage. Their nozzles are designed as horizontally flat orifices, with the spray direction pointing diagonally downwards towards the vacuum suction cup foot array on both sides of the fuselage. Each nozzle can be supplied with air from one of the two independent third outputs of the valve assembly. This structure allows the central processing unit to apply asymmetrical thrust to both sides, while also ensuring that the ejected airflow covers the edge area where the suction cup contacts the working surface, performing the purging task.

[0138] The central processing unit is installed in a control box with electromagnetic shielding and is electrically connected to the electromagnetic drive end of the multi-way proportional directional valve assembly. The processor receives real-time data from the IMU, adhesion pressure, fuselage height, and rotor angle through a reserved sensor interface, and calculates the optimal fluid distribution command according to the aforementioned control method.

[0139] Example 7

[0140] Unlike the above-mentioned system, the omnidirectional convergent surface is composed of a flexible covering layer 8a laid on the bottom of the robot chassis 1. The inner cavity of the robot chassis 1 is provided with a surface adjustment mechanism, which includes a displacement driving component and a top support component that is driven by the displacement driving component to perform lifting and lowering movements. The working end of the top support component abuts against the inner surface of the flexible covering layer 8a, and is used to drive the flexible covering layer 8a to deform through the lifting and lowering displacement of the top support component, so as to dynamically adjust the bulge height and curvature of the omnidirectional convergent surface. The surface adjustment mechanism also includes a locking component, which is used to fix or release the top support component at the current lifting and lowering position.

[0141] When implementing the surface adjustment mechanism, refer to the attached document. Figure 4 A simplified diagram of the three-position adjustment mechanism is provided. A rigid internal support frame 5 is horizontally spanned within the inner cavity of the robot chassis. The internal support frame 5 serves as the physical anchor point for the entire adjustment mechanism, and an adjustment motor 6 is mounted on it. The output shaft of the adjustment motor 6 is connected via a coupling or direct drive to an eccentric rounded rectangular cam 7 or an eccentric rounded polygonal cam (the more sides, the more positions). The cam's geometry is a rectangular structure with smooth transition angles, and its rotation axis is offset from the geometric center of the rectangle. The contour surface of the eccentric rounded rectangular cam 7 abuts against the inner surface of the flexible covering layer 8a in real time; preferably, a polymer wear-resistant rubber sleeve 8 is adhered at the contact point on the inner side of the flexible covering layer 8a to buffer the tangential friction generated during cam rotation and protect the structural integrity of the flexible covering layer 8a.

[0142] During the surface adjustment operation, the central processing unit drives the adjustment motor 6 to rotate the eccentric rounded rectangular cam 7 within a preset angle range based on the collected blade surface geometric parameters. When the eccentric rounded rectangular cam 7 rotates to the point where its minor axis points downwards from the chassis, the cam's support lift on the flexible coating layer 8a is at its minimum. As the motor continues to rotate, the cam's profile transitions from the minor axis direction to the major axis direction (or diagonal direction), and its radial displacement acting on the inner side of the flexible coating layer 8a gradually increases, thereby forcing the flexible coating layer 8a to bulge downwards, increasing the bulge height of the omnidirectional convergent surface. This adjustment method based on the change of eccentric profile enables continuous and wide-range adjustment of the flow channel cross-sectional area, ensuring that the Venturi acceleration flow field can generate optimal transient aerodynamic downpressure in different operating ranges.

[0143] To ensure the steady-state performance of the mechanism under severe aerodynamic impact, this design fully utilizes the physical characteristics of the eccentric rounded rectangular cam 7 and the worm gear adjusting motor 6. Because the adjusting motor 6 possesses a natural mechanical self-locking function, combined with the large contact surface of the eccentric rounded rectangular cam 7 at a specific phase, the adjusted shape and position exhibit high compressive stability. Even when severe pressure fluctuations occur in the abdominal flow channel at the moment the de-icing pulse is released, this fluctuation stress cannot reversely drive the cam to undergo phase drift, thus ensuring that the bulge height and curvature of the omnidirectional convergent surface are always locked in the optimal state.

[0144] Example 8

[0145] Based on the above scheme, the robot chassis 1 is also provided with a dynamic center of gravity compensation mechanism in its inner cavity, which includes a longitudinal slide rail 9 arranged in the inner cavity and an active compensation mass block set on the longitudinal slide rail 9; the active compensation mass block is composed of functional components inside the body; during the synchronization of the release of the pulse jet at the first injection end, the active compensation mass block generates a forward compensation displacement toward the front end of the body, and uses the transient inertial torque generated by the forward compensation displacement to neutralize the pitching torque generated by the transient recoil vector on the body.

[0146] The active compensation mass block is connected to the robot chassis 1 via an elastic buffer assembly. When the pulse jet is released at the first injection end, causing the robot body to generate a backward instantaneous acceleration, the active compensation mass block uses physical inertia to compress the elastic buffer assembly and generate a forward compensation displacement. In terms of layout, the inner cavity of the robot chassis 1 is arranged in a double-layer structure. The layer closer to the surface of the rotating blades is the outer layer 200, which houses the dynamic center of gravity compensation mechanism. Conversely, the layer farther from the surface of the rotating blades is the inner layer 100, which houses core components such as the central processing unit, vacuum pump, air compressor, and multi-way proportional reversing valve group. The inner layer 100 is also integrated with the cover of the inner layer 100 along with the outer layer 200. The inner layer 100 and the outer layer 200 are connected by a window on the surface of the outer layer 200 so that the battery pack 10 of the outer layer 200 can be connected to the various electrical components of the inner layer 100 through the window. The side of the outer layer 200 away from the inner layer 100 is the aforementioned support frame 5. The battery pack 10 is connected to the regulating motor 6 through the perforations on the outer layer 200 for power supply.

[0147] A set of longitudinal slide rails 9 are symmetrically installed in the inner cavity of the robot chassis 1 along the longitudinal axis of the body (i.e., parallel to the direction of travel). An active compensation mass block is slidably mounted across the longitudinal slide rails 9. To achieve center of gravity compensation without increasing the load on the robot body, in this embodiment, the originally fixed functional components inside the robot body (preferably a battery pack 10 with a large mass density) are modified into active compensation mass blocks. Elastic buffer components are provided at both ends of the travel of the longitudinal slide rails 9 (at least including the front end). These components consist of a return spring 11 and parallel miniature dampers. The active compensation mass block is connected to the frame structure of the robot chassis 1 through these elastic buffer components. In the initial state, the active compensation mass block is in a preset equilibrium position on the slide rail under the constraint of the spring force.

[0148] When the high-pressure pulse jet is released from the first jet end at the front of the robot body, according to Newton's third law, the robot body will experience a transient backward recoil vector. Due to the robot's lightweight architecture, this recoil force will cause the robot body to generate an instantaneous backward acceleration. At this time, the battery pack 10, acting as a mass block, tends to maintain its original spatial position due to physical inertia, thus generating a forward compensating displacement relative to the backward-moving robot body. During this displacement, the battery pack 10 compresses the return spring 11 (elastic buffer assembly) located at the front end of the longitudinal slide rail 9.

[0149] The compressed front spring instantaneously generates a reaction force pointing forward of the fuselage. This is because the mounting plane of the longitudinal slide rail 9 has a specific lever arm height h relative to the potential overturning axis of the fuselage. r The transient inertial torque M generated during this forward thrust process ine Satisfying the formula:

[0150] M ine =m mass xa mass xh r

[0151] Where, m mass For the mass of battery pack 10, a mass This is the forward acceleration excited by inertia. The direction of this torque is downward pressing against the front of the fuselage, which can neutralize or reduce the pitching torque caused by the pulse jet that causes the fuselage to tilt. This compensation mechanism based on physical inertia has zero delay characteristics and can occur synchronously with the recoil force of the pulse jet and automatically counteract it. In this way, at the moment of pulse jet execution, the system achieves near dynamic balance through passive inertial feedback.

[0152] Then, the system achieves steady-state recovery through the damping characteristics of the elastic buffer component. Within a very short period after the pulse jet ends, the compressed reset spring 11 releases its elastic potential energy, pushing the battery pack 10 back towards the slide rail. At this time, the parallel damper dissipates kinetic energy, suppressing the simple harmonic vibration of the battery pack 10 on the slide rail, allowing it to smoothly return to its initial equilibrium position, thus avoiding the impact of secondary swaying of the mass block on the adhesion stability of the suction cup foot.

[0153] Example 9

[0154] A rotary blade de-icing robot includes the de-icing spray system described in the above embodiment, an attachment drive component, and a power and medium supply component. The attachment drive component is symmetrically arrayed on both sides of the robot chassis 1, including multiple independently controlled vacuum adsorption feet, and each vacuum adsorption foot is provided with an active drive joint for performing step displacement. The active drive joint is electrically connected to a central processing unit. The power and medium supply component includes a supply base station set independently of the robot body, and an umbilical cable connecting the supply base station and the robot body. The supply base station integrates a high-pressure air source, and the umbilical cable integrates a high-pressure air supply hose. One end of the high-pressure air supply hose is connected to the high-pressure air source of the supply base station, and the other end is connected to a high-pressure fluid inlet on the robot body. In this way, the robot body is lightweight by externalizing the base station.

Claims

1. A de-icing jet control method for a rotating blade de-icing robot, acting on an attachment unit formed by a single vacuum suction cup foot or multiple vacuum suction cup foot groups to move on a rotating blade, characterized in that, The robot is equipped with a fluid distribution component. The input end of the fluid distribution component is connected to a high-pressure fluid source, and its output end is at least branched into a first spray end toward the de-icing operation path and a second spray end toward the converging curved surface at the bottom of the robot body. The control method includes: The system acquires the combined external load vector generated by the robot under forces in a rotating reference frame in real time. The combined external load vector is at least composed of the robot's gravity, environmental aerodynamic forces, and rotational inertial forces coupled together. Based on the parameters of the de-icing jet to be executed, the system estimates the transient recoil vector generated at the moment of triggering the first jet. Based on the spatial vector relationship between the integrated external load vector and the transient recoil vector, the jet phase window that satisfies the preset threshold for the fuselage adhesion torque margin is calculated. If it is determined that the current jet phase window is in effect and de-icing operation is triggered, the fluid distribution component is controlled to perform transient multi-channel fluid distribution: a first proportion of high-pressure fluid is directed to the first jet end to output an ice-breaking jet, and a second proportion of high-pressure fluid is simultaneously directed to the second jet end to drive the high-pressure fluid to sweep tangentially along the convergent curved surface at the bottom of the fuselage, so as to generate transient aerodynamic downforce pointing towards the working surface.

2. The de-icing jet control method for a rotating blade de-icing robot according to claim 1, characterized in that, The step of calculating the jet phase window where the fuselage adhesion torque margin meets a preset threshold based on the spatial vector relationship between the integrated external load vector and the transient recoil vector includes the following steps: When the transient recoil vector acts on the fuselage, the potential overturning axis that causes the fuselage to tilt relative to the working surface is determined. Using the potential overturning axis as a reference, calculate the sequence of integrated external load moments generated by the dynamic mapping of the integrated external load vector in the 360-degree full phase domain of the wind turbine rotor rotation, and calculate the recoil overturning moment generated by the transient recoil vector relative to the potential overturning axis. Traverse the sequence of combined external load torques and filter out the phase intervals that make the combined external load torques manifest as stable clamping torques pointing towards the working surface; Within the phase interval, if the sum of the stable clamping torque and the effective adsorption torque provided by the current robot's attachment unit, minus the recoil and overturning torque, results in a residual constraint torque greater than a preset safety reserve threshold, then the corresponding continuous phase interval is marked as the jet phase window.

3. The de-icing jet control method for a rotating blade de-icing robot according to claim 1, characterized in that, The step of controlling the fluid distribution component to perform transient multi-channel fluid distribution and dynamically adjusting the first ratio and the second ratio includes: Obtain the target pressure of the main jet required for the current de-icing operation, and calculate the transient recoil torque generated at the moment of triggering based on the target pressure of the main jet and the nozzle cross-sectional area of ​​the first jet end. Based on the torque balance constraint, the target aerodynamic downforce required to counteract the transient recoil torque is calculated. The preset hydrodynamic acceleration mapping model of the convergent surface at the bottom of the fuselage is invoked to calculate the tangential fluid target velocity required to generate the target aerodynamic downpressure within the convergent surface. Based on the fluid consumption corresponding to the main jet target pressure and the fluid consumption corresponding to the tangential fluid target velocity, the dynamic allocation commands of the first ratio and the second ratio are calculated and output to the drive control terminal of the fluid allocation component. The greater the target pressure of the main jet, the greater the second proportion of fluid kinetic energy controlled and distributed to the second injection end, so that the transient aerodynamic downpressure generated at the bottom of the fuselage can be increased in accordance with the intensity of the ice-breaking jet at the first injection end.

4. The de-icing jet control method for a rotating blade de-icing robot according to claim 3, characterized in that, The control method further includes an adaptive closed-loop correction step for the transient aerodynamic pressure: The attachment status parameters fed back by each attachment unit are monitored in real time, as well as the displacement deviation value of the machine body relative to the working surface is monitored. The adhesion performance loss characteristics are extracted by the adhesion state parameters or the displacement deviation value, and then coupled with the target aerodynamic downpressure to extract the downpressure compensation deviation value. An adaptive correction factor is generated based on the downpressure compensation deviation value, and the adaptive correction factor is used to reconstruct the fluid dynamics acceleration mapping model in real time. The second ratio is dynamically updated based on the reconstructed fluid dynamics acceleration mapping model.

5. The de-icing jet control method for a rotating blade de-icing robot according to claim 3 or 4, characterized in that, The control method further includes a dynamic compensation step for the atmospheric physical environment, including: Real-time acquisition of atmospheric pressure and ambient temperature of the working environment; calculation of air density correction factor under current altitude and meteorological conditions. The parameters of the hydrodynamic acceleration mapping model are calibrated using the air density correction coefficient, so as to dynamically correct the mapping function between the tangential fluid target velocity and the transient aerodynamic pressure according to the changes in the atmospheric physical environment; The second ratio is updated according to the modified mapping function.

6. The de-icing jet control method for a rotating blade de-icing robot according to claim 1, characterized in that, The output end of the fluid distribution component further branches into third injection ends pointing to the attachment units on the left and right sides of the fuselage, respectively; the control method also includes a lateral correction step for smoothing yaw slip, including: The rotational angular velocity of the wind turbine rotor and the displacement velocity of the robot are acquired in real time. The Coriolis force component generated by their coupling is calculated, and the transient yaw disturbance torque is synthesized by combining the lateral eccentricity of the transient recoil vector. Based on the spatial deflection direction and amplitude of the transient yaw disturbance torque, the fluid distribution assembly is controlled to perform asymmetric differential fluid distribution to the third injection ends located on the left and right sides of the fuselage. The third jet end, which is on the same side as the transient yaw disturbance torque deflection trend, is controlled to instantaneously increase the split ratio and output a lateral pressurized jet. The reaction force generated by the lateral pressurized jet is used to construct a lateral correction torque that dynamically counteracts the transient yaw disturbance torque.

7. The de-icing jet control method for a rotating blade de-icing robot according to claim 1, characterized in that, The control method further includes: Real-time monitoring of the operating phase of each attachment unit of the robot is used to determine whether the robot body is currently in a stepping phase where some attachment units are detached from the working surface, or in a support phase where all attachment units are attached to the working surface. Establish an interlock mechanism between the jet trigger command for de-icing operations and the operating phase: If it is determined that the current phase is the step phase, the jet trigger command is suspended, and the fluid distribution component is restricted from performing transient multi-channel fluid distribution until the fuselage returns to the support phase; If it is determined that the current support phase is being reached and de-icing is being triggered, during the synchronization of the fluid distribution component performing transient multi-channel fluid distribution, the drive joints of each attachment unit are controlled to enter a rigid lock-up mode. By utilizing the attachment unit that has entered the rigid locking mode, a rigid transmission path is constructed to allow the impact load caused by the transient recoil vector to penetrate the fuselage and be released to the working surface.

8. The de-icing jet control method for a rotating blade de-icing robot according to claim 1, characterized in that, The control method further includes dynamic protection steps based on the coupling of thermal load and aerodynamic pressure, including: The operating environment temperature and the status of the machine body system are acquired in real time. The system status includes at least the drive joint resistance and edge sealing margin of each attachment unit. If the ambient temperature is determined to be lower than a preset temperature threshold, the fluid distribution component is controlled to guide the thermo-pressurized fluid generated by the high-pressure fluid source to perform anti-icing and heat maintenance operations. The thermo-pressed fluid is directed in a set proportion to the edge guide slots of the drive joint and the attachment unit; If it is determined that the combined external load vector has reached the instability boundary during the de-icing operation, an energy redistribution action is performed: the system is forcibly switched to the second distribution ratio, and the full amount of the thermo-pressurized fluid is directed to the second injection end.

9. A de-icing jetting system for a rotary blade de-icing robot, characterized in that, The de-icing and blowing system includes: A fluid control hub, the fluid control hub including a high-pressure fluid inlet and a multi-way proportional directional valve assembly connected to the high-pressure fluid inlet; The robot chassis has an omnidirectional convergent surface on its belly. The omnidirectional convergent surface rises smoothly inward from the peripheral edge of the robot chassis, and the geometric extreme point of the omnidirectional convergent surface is located in the central region of the belly of the chassis. A nozzle array, distributed at different phases of the fuselage, including: The main working nozzle is located at the front end of the machine body and points to the working path through the first output end of the multi-way proportional reversing valve group; A negative pressure induction nozzle is arranged in the front end area of ​​the body, with its nozzle pointing to the geometric extreme point of the omnidirectional convergent surface, and the spray axis of the negative pressure induction nozzle is parallel to the tangent of the omnidirectional convergent surface on the spray path. Lateral adjustment nozzles are symmetrically arranged on the left and right sides of the machine body and point to the attachment units on both sides of the machine body respectively. The two lateral adjustment nozzles are supplied with air through the two independent third output terminals of the multi-way proportional reversing valve group. The central processing unit is electrically connected to the multi-way proportional directional valve group and is used to execute the de-icing spray control method as described in any one of claims 1 to 8.

10. The de-icing jetting system for a rotary blade de-icing robot according to claim 9, characterized in that: The omnidirectional convergent surface is composed of a flexible covering layer laid on the bottom of the robot chassis; The robot chassis has a shape adjustment mechanism in its inner cavity. The shape adjustment mechanism includes a displacement drive component and a top support component that is driven by the displacement drive component to perform lifting and lowering movements. The working end of the top support abuts against the inner side of the flexible covering layer, and is used to drive the flexible covering layer to deform through the lifting displacement of the top support, so as to dynamically adjust the bulge height and curvature of the omnidirectional convergent surface. The shape adjustment mechanism also includes a locking component for fixing or releasing the top support member at its current lifting position.

Images