High-voltage line deicing robot and deicing method based on active liquid centroid adjustment

Abstract

The application discloses a high-voltage line deicing technology field based on active liquid centroid adjustment, and relates to a high-voltage line deicing robot and a deicing method.The deicing robot comprises a shell, mounting plates arranged on both sides of the shell, an ice-breaking module for deicing, and a self-adaptive walking mechanism for walking along the high-voltage line, and further comprises a liquid balance module for adjusting the centroid of the robot when the stability of the robot is affected, which comprises an IMU sensor, at least one water tank arranged on each mounting plate, and a pumping unit connected between the two water tanks.Through the liquid balance module, a connected fluid circuit is constructed inside the module, the host computer solves the IMU attitude, and the bidirectional pump in the control loop cooperatively transfers the liquid between the two liquid storage cavities quickly.This active mass redistribution mechanism changes the overall centroid position of the robot, thereby generating a reverse restoring torque to offset the tilt and ensure high-precision attitude positioning of the system.

Description

Technical Field

[0001] This invention relates to the field of high-voltage power line de-icing, and specifically to a high-voltage power line de-icing robot and de-icing method based on active liquid centroid adjustment. Background Technology

[0002] Currently, de-icing of high-voltage transmission lines mainly employs three categories of solutions: manual de-icing, DC de-icing, and mechanical de-icing robots. Manual de-icing relies primarily on workers climbing towers and using insulated poles to strike the ice. DC de-icing melts the ice layer by applying a short-circuit current to the line, utilizing the thermal effect. In the field of mechanical automation, existing de-icing robots are typically equipped with rotary cutting blades or high-frequency vibratory hammers. They move along the line via a motor-driven walking mechanism, removing the ice through cutting or vibration. To maintain their upright position in the air, these robots usually have a fixed rigid weight suspended below them as a counterweight, allowing gravity to keep the robot hanging naturally and maintaining its upright position.

[0003] Current mainstream de-icing technologies all have significant objective drawbacks and cannot meet the operational needs under complex weather conditions. First, manual de-icing is extremely inefficient and carries a very high risk of operation at height; DC de-icing consumes a huge amount of energy and requires line shutdown, resulting in high implementation costs. Second, existing mechanical de-icing robots have three major defects: one is poor posture stability; their "rigid passive counterweight" method cannot cope with the center of gravity shift caused by strong crosswinds or severe icing on one side of the conductor, and the robot is prone to swinging significantly or even flipping and derailing under wind force; the other is high risk of damage; the rotary cutting blades need to apply enormous pressure when facing hard ice layers, which can easily cut the aluminum strands of the conductor, causing permanent damage. Summary of the Invention

[0004] The purpose of this invention is to provide a high-voltage line de-icing robot and de-icing method based on active liquid centroid adjustment, which solves the problem of attitude stability of existing de-icing robots.

[0005] The present invention achieves the above objectives through the following technical solutions:

[0006] A high-voltage line de-icing robot based on active liquid center of mass adjustment includes a shell, mounting plates on both sides of the shell, an ice-breaking module for de-icing, an adaptive walking mechanism for walking along the high-voltage line, and a liquid balance module. The liquid balance module is used to adjust the robot's center of mass when the robot's stability is affected. It includes an IMU sensor, at least one water tank on each side mounting plate, and a pumping unit connecting the two water tanks.

[0007] When the robot walks along the high-voltage line, the liquid balance module acquires the robot's tilt angle and angular velocity in real time through the IMU sensor, establishes a torque-liquid transfer mapping model, determines the initial liquid transfer amount based on the tilt angle and torque balance principle, and then compensates for the initial transfer amount based on the angular velocity to determine the final transfer amount, which is then executed by the pumping unit.

[0008] As a preferred embodiment of the present invention, the ice-breaking module includes a motor, a crank connected to the output shaft of the motor, an impact hammer connected to the crank, and a multi-source sensor.

[0009] As a preferred embodiment of the present invention, the determination of the initial liquid transfer amount based on the tilt angle and the torque balance principle includes:

[0010] Calculate the instability moment of the robot caused by crosswind based on the tilt angle. , is represented as:

[0011] ;

[0012] In the formula, For the total mass of the robot, Let be the vertical distance from the robot's center of mass to the point where the wire is suspended. The center-to-center distance between the two water tanks. The tilt angle of the robot;

[0013] Calculate the restoring torque obtained based on the liquid transfer from the water tank. , is represented as:

[0014] ;

[0015] In the formula, For the density of the liquid, This refers to the pumping volume and direction. It is gravitational acceleration;

[0016] The initial transfer amount is obtained based on the balance between the instability moment and the restoring moment, and is expressed as:

[0017] .

[0018] As a preferred embodiment of the present invention, the compensation for the initial transfer amount based on angular velocity includes: obtaining a compensation term based on angular velocity. , Angular velocity, The dynamic damping coefficient is given, and the torque-fluid transfer mapping model is determined, expressed as:

[0019] ;

[0020] In the formula, These are system characteristic parameters. This indicates the final amount transferred.

[0021] As a preferred embodiment of the present invention, the mapping model further includes, in determining the final transfer amount,: [further details regarding the mapping model and its implications]. Real-time online identification and correction are performed, with data collected at preset intervals. ,in, This is the tangent of the robot's current tilt angle. The actual volume of liquid transferred as accumulated by the flow meter;

[0022] Constructing observation equations , bring in , , The estimated value is updated in real time using recursive least squares with a forgetting factor. , with estimated value As a characteristic parameter of the real-time system.

[0023] As a preferred embodiment of the present invention, the ice-breaking process of the ice-breaking module includes:

[0024] Establishing normal contact force during impact With depth of intrusion The differential equation is:

[0025] ;

[0026] In the formula, The equivalent contact stiffness to be identified, The nonlinear damping coefficient to be identified is... The Hertzian contact coefficient is obtained empirically. Impact velocity;

[0027] Calculate the normal contact force during impact based on multi-source sensor data. With depth of intrusion :

[0028] ;

[0029] In the formula, The length of the impact hammer head connecting handle, For motor stator shaft current, The torque constant of the motor. For equivalent rotational inertia, The coefficient of adhesion friction, To determine the real-time swing angle of the impact hammer, To determine the angular velocity of the impact hammer, The angular acceleration of the impact hammerhead, The initial angle of the contact point;

[0030] The differential equation is solved using a parameter identification algorithm based on the discrete time domain, and the current ice layer stiffness is output. and nonlinear damping coefficient The pre-set impact pattern is matched and executed based on the current ice layer stiffness and nonlinear damping coefficient.

[0031] As a preferred embodiment of the present invention, the preset striking mode includes:

[0032] First mode: In When the value exceeds the set threshold, the natural frequency of the ice-hammer coupling system is calculated, and a sinusoidal excitation trajectory is generated so that the impact hammer head excites the ice layer with equal amplitude at the system's natural frequency. The resonance amplification effect causes the internal stress of the ice layer to exceed the fracture limit.

[0033] Second mode: In When the pressure exceeds the set threshold, the impact hammer head maintains continuous static pressure after contacting the ice layer, and generates tangential shear force in conjunction with the movement of the adaptive walking mechanism to peel off the highly adhesive ice layer.

[0034] Third Mode: In and All values ​​are less than the set threshold, and the multi-source sensors detect valid normal contact force signals. At that time, the impact hammer head is superimposed with a high-frequency micro-amplitude signal to form a multi-harmonic micro-pulse, which quickly shatters the residual ice slag.

[0035] As a preferred embodiment of the present invention, the ice-breaking process further includes: acquiring the impact acceleration signal of the impact hammer head based on multi-source sensors, and extracting time-frequency domain features of the impact acceleration signal using discrete wavelet transform to obtain the first... High-frequency detail factor of the layer and low-frequency approximation coefficient :

[0036] ;

[0037] In the formula, These are the coefficients of the low-pass and high-pass filters, respectively.

[0038] Calculate the energy ratio of the level 1 detail coefficient to the level 3 approximation coefficient. :

[0039] ;

[0040] Preset safety threshold If the energy ratio The ice-breaking module stops operating when the impact hammer hits the wire after a sampling cycle of 'a', where 'a' is a preset value.

[0041] As a preferred embodiment of the present invention, in the process of acquiring the robot's tilt angle and angular velocity in real time through the IMU sensor, the EKF algorithm is used to filter out the torque generated by hitting the ice layer as Gaussian white noise, and the low-frequency attitude trend is separated from the noisy data acquired by the IMU sensor to obtain a smooth tilt angle and angular velocity.

[0042] In a preferred embodiment of the present invention, during the process of executing the final transfer amount through the pumping unit, a variable universe fuzzy PID algorithm is used to solve for the target flow rate.

[0043] attitude error and error change rate Mapping to the fuzzy universe, we define the fuzzy subset as Where NB represents negative large, NM represents negative medium, NS represents negative small, ZO represents zero, PS represents positive small, PM represents positive medium, and PB represents positive large; a scaling factor is introduced. When the error When the error increases, the universe of discourse is automatically compressed; when the error increases... As the value approaches zero, the domain of discourse is expanded; based on a pre-defined expert rule base, the correction values ​​of the three parameters of the PID algorithm are output in real time. And update the three parameters in real time. :

[0044] The instantaneous target flow control quantity is calculated based on the following formula. :

[0045] ;

[0046] The target flow control quantity is converted into a PWM drive signal for the pumping unit. When the voltage is less than the pump's start-up dead zone voltage, the base voltage is added. To ensure linear output of small flow rates; when the tilt angle is detected to cross zero, the pumping unit outputs a short-pulse braking signal in the opposite direction to lock the liquid level using fluid back pressure.

[0047] As a preferred embodiment of the present invention, the adaptive walking mechanism includes a drive wheel and a pneumatic telescopic rod disposed within the housing, a clamping wheel for clamping the high-voltage line from both sides, a translation seat disposed on the clamping wheel, a pair of parallel first connecting rods disposed between the translation seat and the housing, and a second connecting rod disposed between the translation seat and the output end of the pneumatic telescopic rod.

[0048] As a preferred embodiment of the present invention, the robot further includes a spraying module, which includes an antifreeze storage tank disposed within the housing and spray nozzles disposed below the housing with adjustable angle.

[0049] In a preferred embodiment of the present invention, the spray nozzle angle is adjustable, and the spray module quantifies airflow disturbance using a pre-established aerodynamic drag model, wherein the drag model is expressed as:

[0050] ;

[0051] In the formula, The lateral wind resistance experienced by the spray droplet jets. Let be the drag coefficient of the droplet swarm. air density, The effective windward cross-sectional area of ​​the atomized droplet beam. The lateral wind speed is perpendicular to the direction of travel; the required deflection angle of the spray nozzles is determined using the principle of vector composition. :

[0052] ;

[0053] In the formula, Let be the initial momentum force of the jet at the nozzle exit. ,in This refers to the mass flow rate of the antifreeze. The initial velocity of the jet;

[0054] To make the spray nozzle deflect at an angle This is to actively compensate for wind force.

[0055] To implement the above-mentioned active liquid centroid adjustment high-voltage line de-icing robot, this invention also proposes a de-icing method for the high-voltage line de-icing robot based on any of the above-mentioned active liquid centroid adjustment methods, comprising the following steps:

[0056] The robot is made to walk along the high-voltage line by an adaptive walking mechanism, and ice is removed by the ice-breaking module during the walking process.

[0057] When the robot is affected by crosswinds and generates tilting torque, the tilt angle and angular velocity of the robot are acquired in real time based on the IMU sensor. The final transfer amount of the two water tanks required to restore torque balance is calculated based on the pre-established torque-liquid transfer amount mapping model, and the final transfer amount is executed by the pumping unit.

[0058] The process of establishing the torque-liquid transfer mapping model includes: determining the initial liquid transfer amount based on the tilt angle and the torque balance principle, compensating for the initial transfer amount based on the angular velocity, and determining the final transfer amount.

[0059] The beneficial effects of this invention are as follows:

[0060] Through the liquid balance module, a connected fluid loop is built inside the module. The host computer calculates the attitude through the IMU and coordinates the bidirectional pump in the control loop to quickly transfer the liquid between the two liquid storage chambers. This active mass redistribution mechanism changes the overall center of mass position of the robot, thereby generating a reverse restoring torque to counteract the tilt and ensure the high-precision attitude positioning of the system.

[0061] By setting up an ice-breaking module, the rotational motion of the motor is converted into a high-frequency reciprocating impact motion, which drives a specially designed non-metallic hammer to break the ice in front of the conductor at a fixed point. By combining multi-source sensors to set the motor's drive signal and using algorithms, it is ensured that while breaking the hard frost, the hammer trajectory will not intrude into the safe envelope of the aluminum strands of the conductor. Attached Figure Description

[0062] Figure 1 This is a three-dimensional illustration of the present invention. Figure 1 ;

[0063] Figure 2 This is a three-dimensional illustration of the present invention. Figure 2 ;

[0064] Figure 3 This is the front view of the present invention;

[0065] Figure 4 For the present invention Figure 3 Sectional view along line AA;

[0066] Figure 5 This is a schematic diagram of the liquid balance module and adaptive walking mechanism of the present invention;

[0067] Figure 6 This is a schematic diagram of the present invention after the casing has been removed;

[0068] Figure 7 This is a flowchart of the liquid dynamic balance control process of the present invention;

[0069] In the diagram: 1. Housing; 101. Mounting plate; 2. Liquid balance module; 21. Water tank; 22. Water pipe; 23. Pump unit; 3. Ice breaking module; 31. Motor; 32. Crank; 33. Impact hammer; 4. Adaptive walking mechanism; 41. Drive wheel; 42. Pneumatic telescopic rod; 43. First connecting rod; 44. Second connecting rod; 45. Translation seat; 46. Pressure wheel; 5. Spray module; 51. Storage box; 52. Spray nozzle. Detailed Implementation

[0070] The present application will now be described in further detail with reference to the accompanying drawings. It should be noted that the following specific embodiments are only used to further illustrate the present application and should not be construed as limiting the scope of protection of the present application. Those skilled in the art can make some non-essential improvements and adjustments to the present application based on the above application content.

[0071] like Figure 1-6 As shown, the high-voltage line de-icing robot based on active liquid center of mass adjustment includes a shell 1, mounting plates 101 disposed on both sides of the shell 1, an ice-breaking module 3 for de-icing, an adaptive walking mechanism 4 for walking along the high-voltage line, and a liquid balance module 2. The liquid balance module 2 is used to adjust the robot's center of mass when the robot's stability is affected. It includes an IMU sensor, at least one water tank 21 disposed on each side mounting plate 101, and a pumping unit connecting the two water tanks 21.

[0072] When the robot walks along the high-voltage line, the liquid balance module 2 obtains the robot's tilt angle and angular velocity in real time through the IMU sensor, establishes a torque-liquid transfer mapping model, determines the initial liquid transfer amount based on the tilt angle and torque balance principle, and then compensates the initial transfer amount based on the angular velocity to determine the final transfer amount, which is then executed by the pumping unit.

[0073] In this embodiment, in order to achieve millisecond-level attitude response, a nonlinear torque-fluid transfer mapping model based on the dynamics of a variable mass system was constructed. The establishment and optimization process of this mapping model is as follows:

[0074] First, based on the principle of torque balance, an equation is established. The total mass of the robot (including the weight of the ice) is: The vertical distance (i.e., pendulum length) from the robot's center of mass to the suspension point of the high-voltage line is: The center-to-center distance between the two water tanks 21 is The liquid density is The acceleration due to gravity is When the robot is affected by crosswinds or eccentric loads, it will tilt at an angle. At that time, the resulting instability and overturning moment is:

[0075] ;

[0076] To counteract this torque, a volume of [missing information] is required. The liquid is transferred from one side tank 21 to the other side tank 21, generating a reverse restoring torque. for:

[0077] ;

[0078] According to the torque equilibrium condition You can obtain:

[0079] .

[0080] Preferably, due to the flow resistance characteristics of the liquid in the pipeline and the pump start-up delay, this embodiment introduces a dynamic compensation term. The final torque-fluid transfer mapping model is as follows:

[0081] ;

[0082] in, These are system characteristic parameters. Indicates the final amount transferred. Angular velocity, This is the dynamic damping coefficient (used to suppress overshoot).

[0083] Preferably, due to the total mass during the de-icing process... and center of gravity height As the ice detaches from the ground, changes occur, leading to It is not a constant; it is used to accurately obtain the system characteristic parameters that vary with ice weight. The controller employs recursive least squares (RLS) with a forgetting factor. Real-time online identification and correction are performed, and the specific steps are as follows:

[0084] Construct a regression model and set a time period. The observation equation is Among them, the observed values , representing the cumulative actual volume of liquid transferred by the flow meter, regression vector , representing the current tilt tangent value, the parameter to be identified. ;

[0085] Iterative calculation process: The controller executes the following recursive calculations with a period of 100ms:

[0086] S1: Calculate gain :

[0087] ;

[0088] In the formula, Forgetting factor, take This is used to assign higher weights to the latest data;

[0089] S2: Update parameter estimates :

[0090] ;

[0091] S3: Update covariance :

[0092] ;

[0093] Parameter update: Through the above iterations, the converged system characteristic parameters are output in real time. And substitute it into the feedforward control model to ensure that the total mass With center of gravity Even when drift occurs, the calculated target liquid volume remains accurate.

[0094] When the liquid balance module is working, the IMU sensor calculates the robot's tilt angle and angular velocity relative to the vertical line of gravity in real time. The main controller then executes the PID algorithm to calculate the reverse restoring torque required to restore balance and the corresponding liquid transfer volume. The ethylene glycol aqueous solution in the water tank 21 on the tilted side of the robot is directionally transported to the higher water tank 21 on the other side through the pump group 23. The liquid transfer volume is monitored in real time by the water flow sensor, and the center of mass movement is accurately calculated by combining the liquid density, thereby generating a reverse torque to counteract the tilt. In this embodiment, two sets of water pipes 22 are responsible for unidirectional liquid transfer to ensure efficient directional delivery. The liquid level sensor continuously monitors the changes in liquid level until the robot's posture parameters return to within the safe threshold.

[0095] In this embodiment, the liquid balance module 2 adopts a layered active centroid adjustment mechanism. It achieves precise suppression of high-altitude disturbances through "attitude detection - target torque solution - liquid migration execution - closed-loop correction". This solution first collects the robot's tilt angle, pitch angle and angular velocity changes in real time through the IMU sensor, solves the required balance recovery torque through the internal PID algorithm, and further converts it into the required transfer amount, so that the liquid mass adjustment can maintain a consistent mechanical response effect under different load conditions.

[0096] In this embodiment, to improve the disturbance response speed, the liquid balance module 2 can further adopt a dual-channel compensation strategy: First, the robot overturning speed is determined based on the magnitude of the angular velocity, and the water pump is started in advance with low flow pre-charge to pre-set a portion of the liquid migration amount for feedforward pre-compensation; Second, the actual liquid level height is fed back by the liquid level sensor in the water tank 21 to ensure that the final liquid volume accurately reaches the target value, effectively suppressing the rapid sway caused by wind load, impact vibration and uneven icing, so that the robot can maintain a stable posture when the wind force changes suddenly.

[0097] In this embodiment, for extreme weather conditions, when the IMU sensor detects that the tilt angle exceeds the preset danger threshold (such as 8°-10°), the adaptive walking mechanism 4 is immediately paused, the liquid dynamic balance adjustment process is elevated to the highest priority task, and the liquid is rapidly migrated at the maximum pump speed, so that the robot can recover to the controllable posture range in the shortest possible time.

[0098] In addition, pressure sensors, level sensors, and temperature sensors are arranged in the liquid circuit to form a complete fault monitoring system. This allows the robot to immediately reduce the liquid migration rate and output a fault indicator when abnormalities such as pump blockage, pipeline blockage, liquid freezing, or sudden drop in liquid level occur. If the abnormality persists, the circuit valve will be automatically closed and the power supply to pump group 23 will be cut off to prevent backflow, cavitation, or damage to pump group 23. To adapt to the cold environment, the robot can be equipped with working fluid temperature compensation, which can automatically activate the micro heating film when the outside temperature is below freezing to avoid a sharp increase in fluid viscosity that would slow down the response. Through the comprehensive implementation of the above control mechanisms, the liquid balance module 2 can continuously output stable attitude adjustment capability under the coupling effect of strong crosswinds, uneven icing at height, and conductor vibration, providing key support for the reliable operation of the whole machine on high-altitude power lines.

[0099] Example 2

[0100] like Figure 1-7 As shown, based on Embodiment 1, the ice-breaking module 3 in this embodiment includes a motor 31, a crank 32 connected to the output shaft end of the motor 31, an impact hammer 33 connected to the crank 32, and a multi-source sensor.

[0101] When the robot is de-icing, the motor 31 drives the crank 32 to perform continuous circular rotation. The rotational power of the crank 32 is transmitted to the impact hammer 33. The impact hammer 33 oscillates back and forth within a specific fan-shaped angle, and obtains significant kinetic energy during the oscillation process. It performs high-frequency, fixed-point physical impact crushing on the ice layer on the surface of the high-voltage line, providing reliable physical boundary conditions for the efficient operation of the entire robot system under complex ice conditions.

[0102] In this implementation, regarding the control strategy, to ensure optimal ice-breaking effect under different ice types and thicknesses, and to improve the controllability of the impact process, this embodiment sets up multi-source sensors, including angle sensors and inertial force feedback units arranged at the crank 32 drive shaft and pivot joint respectively, to collect the instantaneous angle, angular velocity and acceleration of the impact hammer head 33 in real time; to evaluate the actual impact energy of each impact and the recoil amplitude of the ice layer, and to establish an impact energy model based on these data, which can determine the degree of ice layer breakage, hardness difference and residual thickness, thereby dynamically adjusting the output strategy. The motor 31 drive end integrates a high-precision Hall position sensor and phase current sampling circuit to calculate joint torque and motion trajectory in real time; the impact hammer head 33 is embedded with a MEMS wideband micro accelerometer (bandwidth 0-2kHz) to collect high-frequency vibration characteristics at the moment of contact.

[0103] This embodiment employs a composite control method based on velocity feedforward and angle closed-loop. The crank 32 automatically allocates acceleration and deceleration phases according to the target energy in each impact cycle, ensuring that the impact hammer 33 maintains the optimal impact speed at the moment of contact and suppresses excessive vibration during the return phase. For "high-resistance ice layers" in the case of hard frost or thick ice, this scheme automatically extends the acceleration time of the impact hammer 33's forward swing based on acceleration feedback to increase peak kinetic energy. For brittle thin ice or snow-clearing ice layers, a high-frequency low-energy mode is adopted, which avoids excessive impact on the conductor by increasing the impact frequency and reducing the energy of a single strike.

[0104] Specifically, to accurately quantify the physical properties of ice, this invention constructs a nonlinear viscoelastic impact dynamics model based on Hunt-Crossley theory that better reflects the brittle fracture characteristics of ice. This model considers the nonlinear relationship between energy dissipation and penetration depth during the impact process, and includes the following steps:

[0105] Define the normal contact force at the instant of impact contact (i.e., the time from when the hammer head contacts the ice layer to when it bounces away, approximately 5-15 ms). With depth of intrusion The differential equation is:

[0106] ;

[0107] in, The equivalent contact stiffness to be identified (reflecting the hardness of the ice layer). The nonlinear damping coefficient to be identified (reflecting the energy dissipation rate of the ice layer under impact). The Hertzian contact coefficient is an empirical value used for impacts on irregular ice surfaces. , The impact velocity is the instantaneous normal velocity of the hammerhead.

[0108] Obtain the input variables required for the model by following these steps:

[0109] ;

[0110] In the formula, To determine the length of the connecting handle of the impact hammer 33, For motor 31 stator shaft current, The torque constant of the motor. For equivalent rotational inertia, The coefficient of adhesion friction, To determine the real-time swing angle of the impact hammer, To determine the angular velocity of the impact hammer, The angular acceleration of the impact hammerhead, The initial angle of the contact point;

[0111] Set the initial angle of the contact point as (By capturing the point of sudden change in current), the penetration depth is determined based on the following formula:

[0112] ;

[0113] Because the impact process is extremely short, in order to achieve real-time calculation, the controller discretizes the above continuous equation into a difference equation and solves it using recursive least squares (RLS) with a forgetting factor:

[0114] Constructing the linear regression form: Let Regression vector The parameter vector to be identified Then at that moment The observation equation is: ;

[0115] Iterative calculation steps: The controller operates at a frequency of 2kHz. Perform the following recursive operations:

[0116] Calculate the gain vector :

[0117] ;

[0118] In the formula, Forgetting factor, take This is used to assign higher weights to the latest data;

[0119] Update parameter estimates :

[0120] ;

[0121] Real-time output of current ice stiffness and nonlinear damping coefficient ;

[0122] Update covariance matrix :

[0123] .

[0124] Using the above algorithm, the robot can achieve algorithm convergence before the end of an impact (for example, there are 10 sampling points during an impact in this embodiment), thereby accurately obtaining the mechanical fingerprint of the ice layer and providing numerical basis for subsequent variable impedance control.

[0125] Adaptive de-icing control strategy based on variable impedance: after identifying the ice layer stiffness and nonlinear damping coefficient Subsequently, the robot's controller adopts a two-degree-of-freedom impedance control algorithm to virtualize the impact hammer 33 as a "mass-spring-damping" system, and adapts to different ice conditions by dynamically adjusting the virtual parameters.

[0126] Construct a controller based on the impedance control law equation, execute the following torque control law, and calculate the required output torque of the motor. :

[0127] ;

[0128] in, This refers to the actual movement state of the hammerhead; For the desired trajectory; This is a feedforward torque based on the dynamic model, used to compensate for the gravity term. and friction term ; The controller adjusts the "virtual stiffness", "virtual damping" and "virtual inertia" in real time.

[0129] Hierarchical adaptive adjustment logic: Based on the identification results, the virtual parameters and desired trajectory in the above equations are corrected in real time. :

[0130] First mode: Resonance-based fracturing for high-rigidity brittle ice (rime);

[0131] when First, calculate the natural frequencies of the ice-hammer coupling system. In the formula The moment of inertia of the impact hammer head; virtual stiffness Lower the virtual damping Set the damping to 0.2 times the critical damping, i.e., the underdamped state; generate a sinusoidal excitation trajectory. The impact hammer 33 excites the ice layer with constant amplitude at the system's natural frequency, utilizing the resonance amplification effect to cause the internal stress of the ice layer to exceed the fracture limit. In the formula... The hammerhead angle at the moment of contact. This represents the excitation amplitude.

[0132] Second mode: constant force shearing for high-viscosity soft ice (mixed frost);

[0133] when At that time, virtual stiffness Adjust to 80% of the maximum value, virtual damping Increase to an overdamped state to prevent rebound; use a step signal. and limit the maximum output torque. , This is the preset push angle increment. The rated torque of the motor ensures that the impact hammer 33 maintains continuous static pressure after contact, which, in conjunction with the movement of the traveling mechanism, generates tangential shearing force to peel off the highly adhesive ice layer.

[0134] Third mode: Micropulse sweeping is used to target residual thin ice;

[0135] exist and All values ​​are less than the aforementioned threshold, and the multi-source sensors detect valid normal contact force signals. Set virtual stiffness (to exclude impact-induced slippage). for Virtual damping Set to zero, so that the impact hammer 33 is superimposed with a high-frequency micro-amplitude signal. In the formula, The current sweeping reference angle of the impact hammer head; For the first Amplitude coefficients of the second harmonic components; The harmonic order is 1, 2, or 3. The fundamental frequency of the micropulse (in this embodiment, it is taken as...) The above equations generate multi-harmonic micropulses, which rapidly shatter residual ice slag.

[0136] Furthermore, in order to distinguish between "accidental conductor strike" and "normal ice impact" within milliseconds, this embodiment employs Discrete Wavelet Transform (DWT) to extract time-frequency domain features from the impact acceleration signal, thus constructing an instantaneous fuse protection mechanism:

[0137] Acquire raw signals from the hammerhead MEMS accelerometer. The Daubechies4 (db4) wavelet basis was selected for 3-level Mallat decomposition. High-frequency detail factor of the layer and low-frequency approximation coefficient The calculation is as follows:

[0138] ;

[0139] in, These are the coefficients of the low-pass and high-pass filters, respectively.

[0140] Calculate the energy ratio of the first-layer detail coefficients (high-frequency components, representing brittle fracture shock waves) to the third-layer approximation coefficients (low-frequency components, representing flexible rebound). :

[0141] ;

[0142] Preset safety threshold (Based on extensive experimental calibration), if the calculation results If two consecutive sampling cycles are triggered, the controller determines that the metal wire has been accidentally struck. At this time, the underlying drive logic directly bypasses the PID control loop, generates the maximum reverse electromagnetic braking torque, and forces the motor to lock the rotor in a short time to prevent the impact hammer 33 from further damaging the high-voltage line.

[0143] Furthermore, in complex ice conditions (such as layered ice, mixed ice, and ice nodules), the robot can incorporate an ice condition adaptive adjustment mechanism. When the rebound acceleration after the impact of the hammer 33 exceeds a preset threshold, it indicates that the ice layer is locally hard. At this time, the torque output of the motor 31 is automatically increased to enhance the energy of the subsequent three to five impacts. If the acceleration after two consecutive impacts is significantly reduced, it is determined that the ice layer has been locally broken. The robot will automatically reduce the impact force and appropriately increase the frequency to accelerate the ice block shedding efficiency.

[0144] To ensure absolute safety of the conductors during the de-icing process, this ice-breaking module 3 is equipped with multiple safety constraints. First, the swing trajectory of the impact hammer 33 is monitored in real time. When the angular displacement exceeds the set safety envelope, the power output is immediately interrupted to prevent the impact hammer 33 from accidentally touching the high-voltage line. Second, the current and operating temperature of the motor 31 are monitored. When an abnormal increase in load or excessive temperature rise of the motor 31 is detected, the impact frequency will be automatically reduced or the system will enter the protection mode. Finally, if no effective ice-breaking feedback is observed after more than a set number of continuous high-energy impacts, the system will automatically assume that there may be non-ice-like hard objects or foreign objects blocking the target area, and will suspend the operation and prompt manual inspection.

[0145] Through the synergistic effect of the above-mentioned multi-dimensional control strategies, the reciprocating impact ice-breaking module 3 can adaptively adjust to different ice layer characteristics, ensuring the dynamic optimal state of impact frequency, impact energy and impact hammer 33 trajectory, effectively improving de-icing efficiency while avoiding damage to the wire structure, and providing key execution capabilities for the safe and stable operation of the whole machine in harsh ice conditions.

[0146] To reduce the impact of the ice-breaking action of the ice-breaking module 3 on the liquid balance module 2, the balance algorithm implemented in this embodiment consists of a three-level processing flow: signal preprocessing layer, fuzzy decision layer, and execution control layer.

[0147] Step 1: State estimation based on extended Kalman filter (EKF) (signal layer) Because the impact frequency of the mechanical de-icing hammer (approximately 3-5Hz) will couple into the IMU data, causing oscillations in the attitude calculation, the state equation is established as follows:

[0148] ;

[0149] Using the EKF algorithm, the de-icing impact torque is treated as Gaussian white noise and filtered out. The real low-frequency attitude trend is separated from the noisy accelerometer and gyroscope data, and the smooth tilt angle and angular velocity are obtained as inputs for subsequent control.

[0150] Step 2: Solve for the target flow rate using a variable universe of discourse fuzzy PID algorithm:

[0151] Blurring: This process reduces attitude error. and error change rate Mapping to the fuzzy universe, we define the fuzzy subset as NB represents negative large, NM represents negative medium, NS represents negative small, ZO represents zero, PS represents positive small, PM represents positive medium, and PB represents positive large;

[0152] Scaling factor adjustment: Introducing a scaling factor When the error When the error increases, the domain of discourse is automatically compressed, which is equivalent to increasing the control gain; when the error... As the domain approaches zero, the domain is expanded, increasing the control resolution.

[0153] Based on the expert rule base of "strong correction for large deviations and gentle adjustment for small deviations", the correction values ​​of the three parameters of the PID algorithm are output in real time. And update three parameters in real time. ;

[0154] Calculate the instantaneous target flow control quantity:

[0155] ;

[0156] Step 3: Calculate the control quantity The target flow rate is converted into a PWM drive signal for the pump motor. This process also includes compensation for the nonlinear characteristics of the pump unit 23.

[0157] Dead zone transition: when When the voltage is less than the pump's start-up dead zone voltage, a base voltage is added. This ensures linear output even with minimal flow rates.

[0158] Reverse braking: When tilt angle is detected At the instant the liquid level crosses zero (i.e., returns to horizontal), the algorithm outputs a short, reverse-pulse braking signal (lasting 50-100ms) to quickly lock the liquid level using fluid back pressure, preventing repeated oscillations caused by fluid inertia.

[0159] Example 3

[0160] like Figure 1-6In this embodiment, the adaptive walking mechanism 4 includes a drive wheel 41 and a pneumatic telescopic rod 42 disposed in the housing 1, a clamping wheel 46 for clamping the high voltage line from both sides, a translation seat 45 disposed on the clamping wheel 46, a pair of parallel first connecting rods 43 disposed between the translation seat 45 and the housing 1, and a second connecting rod 44 disposed between the translation seat 45 and the output end of the pneumatic telescopic rod 42.

[0161] The robot also includes a spray module 5, which includes an antifreeze storage tank 51 disposed inside the housing 1 and an angle-adjustable spray nozzle 52 disposed below the housing 1.

[0162] The adaptive walking mechanism 4 operates as an integrated flexible mechanical control process. Compressed air is injected into the chamber inside the pneumatic telescopic rod 42 to extend axially, forming the main clamping drive circuit. The thrust of the pneumatic telescopic rod 42 is transmitted to the end clamping wheel 46 through the connecting rod, causing it to press tightly against the surface of the high-voltage line at a preset pressure, forming a reliable friction coupling interface. This ensures stable contact between the clamping wheel 46 and the high-voltage line without slippage, and also ensures constant clamping friction, improving the robot's ability to climb steep slopes at high altitudes and resist wind.

[0163] The adaptive walking mechanism 4 utilizes the compressibility of air to form a natural air spring adaptive circuit. When the pressure wheel 46 encounters an ice agglomerate and undergoes a geometrical change, the impact energy is instantly converted into pressure fluctuations within the air chamber, and the pneumatic telescopic rod 42 subsequently generates displacement compensation.

[0164] Through this compliant clamping logic, the mechanism can maintain a constant clamping torque through pneumatic feedback, regardless of slight differences in wire diameter or changes in environmental load, ensuring the robot's safety and operational continuity in all weather conditions and complex working environments.

[0165] In addition, when abnormally low pressure is detected, the load on drive wheel 41 suddenly increases, or the tilt angle exceeds the safety threshold, the system will immediately reduce the traveling speed and increase the clamping force. If the abnormality persists, the pneumatic locking and emergency stop mechanism will be automatically triggered to prevent the robot from derailing or slipping on the guide. Through the above control strategy, the pneumatic adaptive walking mechanism 4 not only achieves stable walking in conventional guide environments, but also has high robustness against wind disturbance, uneven icing, and crossing irregular structures, which significantly improves the robot's engineering applicability in high-altitude real-world scenarios.

[0166] To address the problem of secondary icing on the surface of the conductor after physical de-icing, the robot is equipped with an antifreeze spraying module 5 at its tail. This module is connected to the antifreeze storage tank 51 via an internal pipeline. As the final stage of the operation, it uses fan-shaped atomizing nozzles to evenly cover the clean high-voltage line surface that has just been rolled with hydrophobic solvent. The control system linearly adjusts the spray flow rate of the spraying module 5 according to the robot's walking speed, thereby achieving a uniform film of constant thickness for the antifreeze coating.

[0167] In terms of control strategy, the spray module 5 adopts a spray control model based on the coupling of three parameters: flow rate, pressure and speed. This ensures that a uniform, continuous and firmly adhered antifreeze coating can be formed on the surface of the conductor under different weather conditions. The spray module 5 integrates a micro flow meter, a liquid supply pipeline pressure sensor, a nozzle outlet temperature sensor and a liquid level detection module, thereby constructing a complete closed loop for monitoring the spray process. Before the start of operation, the controller first corrects the droplet evaporation coefficient of the spray area based on the ambient temperature, air humidity and wind speed, and dynamically calculates the target spray volume and spray cycle based on the real-time speed feedback from the walking mechanism.

[0168] During spraying, the controller adjusts the PWM duty cycle of the spray pump and the opening of the nozzle solenoid valve to achieve continuous micro-level output of antifreeze. When the nozzle outlet pressure is detected to be too low, the system automatically increases the pump speed to maintain stable atomization pressure and avoids intermittent spraying or uneven coating thickness. If an abnormal increase in pipeline pressure is detected, it is automatically determined that the nozzle is blocked or the pipeline is frozen. The system will immediately stop spraying and start the nozzle heating unit to raise the local temperature of the spray hole to a safe temperature range in a short time, thereby restoring the spraying capacity.

[0169] To further improve the spraying effect, this embodiment introduces an adaptive adjustment algorithm for the spraying mode. When the wind speed is high or the angle between the wind direction and the direction of travel is large, the spraying device will automatically switch to "lateral compensation mode". By adjusting the deflection angle of the nozzle and the spray pulse frequency, the atomized droplets can still be accurately deposited on the surface of the conductor under the action of wind. When the ambient temperature is too low (such as below -15°C) and the viscosity of the antifreeze increases, the system will automatically extend the pulse width of a single spray and slightly reduce the travel speed to ensure that the antifreeze film reaches the designed thickness. If the robot detects that frost is forming rapidly on the surface or the humidity rises sharply, it will increase the spray volume coefficient in a short time to locally thicken the conductor to delay the freezing process.

[0170] The system has optimized the control logic for continuous light wind environments, enabling the nozzles to perform antifreeze tasks while also having an adaptive balance adjustment function. Through the integrated feedback-feedforward joint control algorithm, it can identify the small tilting torque caused by light wind in real time and adjust the vector direction of the nozzle spray proportionally. It uses the inherent momentum during the spraying process as a compensation power source to achieve uniform coverage of antifreeze.

[0171] Preferably, the spray nozzle 52 has an adjustable angle, and the spray module 5 uses a pre-established aerodynamic drag model to quantify airflow disturbance. The drag model is expressed as follows:

[0172] ;

[0173] In the formula, The lateral wind resistance experienced by the spray droplet jets. Let be the drag coefficient of the droplet swarm. air density, The effective windward cross-sectional area of ​​the atomized droplet beam. The lateral wind speed is perpendicular to the direction of travel; the required deflection angle of the spray nozzle 52 is determined using the principle of vector composition. :

[0174] ;

[0175] In the formula, Let be the initial momentum force of the jet at the nozzle exit. ,in This refers to the mass flow rate of the antifreeze. The initial velocity of the jet;

[0176] To make the spray nozzle 52 deflect at an angle This is to actively compensate for wind force.

[0177] The feedback loop uses high-frequency IMU data to construct a PID closed loop, which corrects model uncertainties and transient pulse winds in real time. The two work together to ensure the robot's posture robustness during anti-freezing spraying operations.

[0178] In terms of safety protection, the spray module 5 can be set to redundancy detection of liquid level. When the antifreeze is insufficient, the controller plans the operation path in advance based on the remaining spraying time, and stops spraying and prompts for replenishment when the liquid level is too low. Low-power heating strips are installed on the outer layer of the antifreeze pipeline and the pump body. In extremely cold environments, they are activated by temperature sensors to prevent the liquid from freezing or excessively increasing in viscosity, which would affect the spraying performance. The system can also be set to set a self-checking mechanism for the spraying area. By comparing the pressure and flow changes before and after the nozzle and the light reflection characteristics of the wire surface (optional vision module), it can determine whether the spraying is covering normally. If insufficient coverage is detected, the forward speed is automatically reduced or the number of sprays is increased to ensure the continuity of the coating.

[0179] Through the above control strategies, the antifreeze spraying device can maintain a stable and reliable spraying effect under adverse conditions such as wind disturbance, low temperature freezing, changes in travel speed and changes in liquid viscosity, and minimize the risk of secondary icing of the conductor, achieving a high degree of synergy between de-icing and anti-icing.

[0180] To implement the above-mentioned active liquid centroid adjustment high-voltage line de-icing robot, this invention also proposes a de-icing method for the high-voltage line de-icing robot based on any of the above-mentioned active liquid centroid adjustment methods, comprising the following steps:

[0181] The robot is made to walk along the high-voltage line by the adaptive walking mechanism 4, and the ice-breaking module 3 is used to remove ice during the walking process.

[0182] When the robot is affected by crosswind and generates tilting torque, the tilt angle and angular velocity of the robot are acquired in real time based on the IMU sensor. The final transfer amount of the two water tanks 21 required to restore torque balance is calculated based on the pre-established torque-liquid transfer amount mapping model, and the final transfer amount is executed by the pumping unit.

[0183] The process of establishing the torque-liquid transfer mapping model includes: determining the initial liquid transfer amount based on the tilt angle and the torque balance principle, compensating for the initial transfer amount based on the angular velocity, and determining the final transfer amount.

[0184] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A high-voltage line de-icing robot based on active liquid centroid adjustment, comprising a housing (1), mounting plates (101) disposed on both sides of the housing (1), an ice-breaking module (3) for de-icing, and an adaptive walking mechanism (4) for walking along the high-voltage line, characterized in that, It also includes a liquid balance module (2) for adjusting the robot's center of mass when the robot's stability is affected. The liquid balance module (2) includes an IMU sensor, at least one water tank (21) disposed on each side mounting plate (101), and a pumping unit connecting the two water tanks (21). When the robot walks along the high-voltage line, the liquid balance module (2) obtains the tilt angle and angular velocity of the robot in real time through the IMU sensor, establishes a torque-liquid transfer mapping model, determines the initial liquid transfer amount based on the tilt angle and torque balance principle, and then compensates the initial transfer amount based on the angular velocity to determine the final transfer amount, and executes the final transfer amount through the pumping unit. The ice-breaking module (3) includes a motor (31), a crank (32) connected to the output shaft end of the motor (31), an impact hammer (33) connected to the crank (32), and a multi-source sensor; The ice-breaking process of the ice-breaking module (3) includes: Establishing normal contact force during impact With depth of intrusion The differential equation is: ； In the formula, The equivalent contact stiffness to be identified, The nonlinear damping coefficient to be identified is... The Hertzian contact coefficient is obtained empirically. Impact velocity; Calculate the normal contact force during impact based on multi-source sensor data. With depth of intrusion : ； ； In the formula, The length of the connecting handle of the impact hammer (33) For motor stator shaft current, The torque constant of the motor. For equivalent rotational inertia, The coefficient of adhesion friction, To determine the real-time swing angle of the impact hammer, To determine the angular velocity of the impact hammer, The angular acceleration of the impact hammerhead, The initial angle of the contact point; The differential equation is solved using a parameter identification algorithm based on the discrete time domain, and the current ice layer stiffness is output. and nonlinear damping coefficient The pre-set impact pattern is matched and executed based on the current ice layer stiffness and nonlinear damping coefficient.

2. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 1, characterized in that, The determination of the initial liquid transfer volume based on the tilt angle and torque balance principle includes: Calculate the instability moment of the robot caused by crosswind based on the tilt angle. , is represented as: ； In the formula, For the total mass of the robot, Let be the vertical distance from the robot's center of mass to the point where the wire is suspended. The center-to-center distance between the two water tanks (21) is... The tilt angle of the robot; Calculate the restoring torque obtained from the liquid transfer in the water tank (21). , is represented as: ； In the formula, For the density of the liquid, This refers to the pumping volume and direction. It is gravitational acceleration; The initial transfer amount is obtained based on the balance between the instability moment and the restoring moment, and is expressed as: 。 3. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 2, characterized in that, The process of compensating for the initial transfer amount based on angular velocity to determine the final transfer amount includes: obtaining the compensation term based on angular velocity. , Angular velocity, The dynamic damping coefficient is given, and the torque-fluid transfer mapping model is determined, expressed as: ； In the formula, These are system characteristic parameters. This indicates the final amount transferred.

4. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 3, characterized in that, The mapping model, in determining the final transfer amount, also includes: [further details needed]. Real-time online identification and correction are performed, with data collected at preset intervals. ,in, This is the tangent of the robot's current tilt angle. The actual volume of liquid transferred as accumulated by the flow meter; Constructing observation equations , bring in , The estimated value is updated in real time using recursive least squares with a forgetting factor. , with estimated value As a characteristic parameter of the real-time system.

5. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 1, characterized in that, The preset striking patterns include: First mode: In When the value is greater than the set threshold, the natural frequency of the ice-hammer coupling system is calculated, and a sinusoidal excitation trajectory is generated so that the impact hammer (33) excites the ice layer with equal amplitude at the system's natural frequency. The resonance amplification effect causes the internal stress of the ice layer to exceed the fracture limit. Second mode: In When the impact hammer (33) is greater than the set threshold, it maintains a continuous static pressure after contacting the ice layer, and moves in conjunction with the adaptive walking mechanism (4) to generate tangential shear force to peel off the highly adhesive ice layer. Third Mode: In and All values ​​are less than the set threshold, and the multi-source sensors detect valid normal contact force signals. At that time, the impact hammer (33) is superimposed with a high-frequency micro-amplitude signal to form a multi-harmonic micro-pulse, which quickly shatters the residual ice slag.

6. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 5, characterized in that, The ice-breaking process also includes: acquiring the impact acceleration signal of the impact hammer (33) based on multi-source sensors, and extracting the time-frequency domain features of the impact acceleration signal using discrete wavelet transform to obtain the first... High-frequency detail factor of the layer and low-frequency approximation coefficient : ； In the formula, These are the coefficients of the low-pass and high-pass filters, respectively. Calculate the energy ratio of the level 1 detail coefficient to the level 3 approximation coefficient. : ； Preset safety threshold If the energy ratio If a sampling cycle is triggered continuously, where a is a preset value, it is determined that the impact hammer (33) hits the wire, and the ice-breaking module (3) stops running.

7. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 1, characterized in that, In the process of acquiring the robot's tilt angle and angular velocity in real time through the IMU sensor, the EKF algorithm is used to filter out the torque generated by hitting the ice layer as Gaussian white noise. The low-frequency attitude trend is separated from the noisy data acquired by the IMU sensor to obtain a smooth tilt angle and angular velocity.

8. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 7, characterized in that, During the process of executing the final transfer volume through the pumping unit, a variable universe of discourse fuzzy PID algorithm is used to solve for the target flow rate. attitude error and error change rate Mapping to the fuzzy universe, we define the fuzzy subset as Where NB represents negative large, NM represents negative medium, NS represents negative small, ZO represents zero, PS represents positive small, PM represents positive medium, and PB represents positive large; a scaling factor is introduced. When the error When the error increases, the universe of discourse is automatically compressed; when the error increases... As it approaches zero, the domain of discourse is expanded; Based on a pre-defined expert rule base, the correction values ​​for the three parameters of the PID algorithm are output in real time. And update the three parameters in real time. : The instantaneous target flow control quantity is calculated based on the following formula. : ； The target flow control quantity is converted into a PWM drive signal for the pumping unit. When the voltage is less than the pump's start-up dead zone voltage, the base voltage is added. To ensure linear output of small flow rates; when the tilt angle is detected to cross zero, the pumping unit outputs a short-pulse braking signal in the opposite direction to lock the liquid level using fluid back pressure.

9. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 1, characterized in that, The adaptive walking mechanism (4) includes a drive wheel (41) and a pneumatic telescopic rod (42) installed in the housing (1), and a clamping wheel (46) for clamping the high voltage line from both sides. A translation seat (45) is provided on the clamping wheel (46). A pair of parallel first connecting rods (43) are provided between the translation seat (45) and the housing (1). A second connecting rod (44) is also provided between the translation seat (45) and the output end of the pneumatic telescopic rod (42).

10. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 1, characterized in that, The robot also includes a spray module (5), which includes an antifreeze storage tank (51) disposed in the housing (1) and an angle-adjustable spray nozzle (52) disposed below the housing (1).

11. The high-voltage line de-icing robot based on active liquid centroid adjustment according to claim 10, characterized in that, The angle of the spray nozzle (52) is adjustable, and the spray module (5) quantifies the airflow disturbance through a pre-established aerodynamic drag model, which is expressed as: ； In the formula, The lateral wind resistance experienced by the spray droplet jets. Let be the drag coefficient of the droplet swarm. air density, The effective windward cross-sectional area of ​​the atomized droplet beam. The lateral wind speed is perpendicular to the direction of travel; the required deflection angle of the spray nozzle (52) is determined using the principle of vector synthesis. : ； In the formula, Let be the initial momentum force of the jet at the nozzle exit. ,in This refers to the mass flow rate of the antifreeze. The initial velocity of the jet; The spray nozzle (52) is deflected at an angle. This is to actively compensate for wind force.

12. A de-icing method based on the high-voltage line de-icing robot with active liquid centroid adjustment as described in any one of claims 1-11, characterized in that, Includes the following steps: The robot walks along the high-voltage line through the adaptive walking mechanism (4), and the ice is removed by the ice-breaking module (3) during the walking process; When the robot is affected by crosswind and generates tilting torque, the tilting angle and angular velocity of the robot are obtained in real time based on the IMU sensor. The final transfer amount of the two water tanks (21) required to restore torque balance is calculated based on the pre-established torque-liquid transfer amount mapping model, and the final transfer amount is executed by the pumping unit. The process of establishing the torque-liquid transfer mapping model includes: determining the initial liquid transfer amount based on the tilt angle and the torque balance principle, compensating for the initial transfer amount based on the angular velocity, and determining the final transfer amount.

Images