A training system and method for hydraulic fastening of wind turbine bolts based on GWO standards

Abstract

This invention relates to the field of wind power operation and maintenance training technology, and discloses a wind power bolt hydraulic tightening training system and method based on the GWO standard. The method collects the fingertip pressure distribution matrix, head posture data and breathing frequency of the operator, calculates the virtual fluid injection effectiveness index, and generates a virtual fluid dynamic liquid surface function and liquid surface fluctuation rate. Based on this, it performs virtual breathing light strip rendering, generates operation background sound and provides visual warnings, and calculates the virtual fluid viscosity coefficient and converts it into damping control current to realize variable damping tactile feedback and tactile pre-compression tightening and reverse ejection bidirectional interaction. This invention establishes a tactile communication channel that bypasses language barriers in visual isolation and high-noise environments by constructing a virtual fluid model that couples biological characteristics and mechanical movements. By utilizing flexible damping interlocking and bidirectional tactile interaction, it effectively solves the technical problems of low coordination and high safety hazards caused by lack of perception in collaborative operations in confined spaces.

Description

Technical Field

[0001] This invention relates to the field of technology, and more specifically, to a training system and method for hydraulic fastening of wind turbine bolts based on the GWO standard. Background Technology

[0002] The operation and maintenance of wind power equipment is a crucial link in ensuring the safe operation of the unit. With the continuous increase in unit capacity, the internal structure of the wind turbine hub is becoming increasingly complex, and the technological requirements for critical operations such as bolt tightening are becoming increasingly stringent. When performing high-strength bolt hydraulic tightening operations, two operators typically need to work closely together: one person enters the confined space inside the hub to position and support the hydraulic wrench (operator), while the other person is positioned outside the hub or on the tower platform to operate the hydraulic pump station (pump operator). Due to the special working environment, there is physical isolation between the operator and the pump operator, accompanied by high-intensity mechanical noise. Traditional voice communication methods often cannot guarantee the real-time and accurate transmission of information, posing a significant challenge to the level of coordination and understanding between the two operators.

[0003] In the prior art, there have been some studies on the detection and control of hydraulic equipment or fasteners. For example, Chinese patent CN119783587B discloses a method for characterizing the fluid flow pulsation characteristics of a hydraulic pump based on digital twin technology. This method establishes a theoretical structural model of the hydraulic pump and performs simulation optimization by combining measured data, achieving accurate characterization of the outlet fluid pressure pulsation characteristics of the hydraulic pump, providing an effective means for the stability analysis of hydraulic systems. Another example is Chinese patent application CN117147041A, which discloses an ultrasonic stress detection method for high-strength bolts used in wind turbines. This method utilizes ultrasonic time-of-flight data combined with a neural network model to achieve online detection of bolt stress levels, improving the reliability of bolt connection status assessment. The aforementioned prior art mainly focuses on the performance characterization or status detection of a single device (hydraulic pump or bolt), providing a foundation for understanding the physical characteristics of the equipment.

[0004] However, in wind power operation and maintenance practical teaching and training scenarios involving multiple people, while the aforementioned technologies can provide data support at the equipment level, they cannot solve the most critical problem of "non-line-of-sight collaborative perception" in confined space operations. Specifically, in real-time coordination inside and outside the hub, the pump station operator cannot directly observe the details of the operator's movements, nor can the operator anticipate the pump station operator's pressurization intentions. Existing training methods often rely on walkie-talkie commands or simple signal light indications. This information transmission method based on discrete symbols suffers from severe lag and ambiguity, failing to convey the subtle physiological-mechanical coupling information contained in the operator's "ready" state, such as muscle tension stability and respiratory rhythm regularity. This perceptual gap prevents trainees from establishing intuitive, tacit cooperation when facing high-pressure environments in real operations (such as noise interference and visual obstruction). For example, when a pump operator pressurizes solely based on verbal commands, if the operator's hand trembles slightly or their grip posture is slightly adjusted due to nervousness, the operator may not be able to immediately perceive and stop the operation, easily leading to safety accidents such as the wrench slipping and trapping the operator's hand. Conversely, if the operator has completed preparations but delays pressurization due to hesitation, it will result in low work efficiency. Existing teaching methods fail to transform this implicit physiological-psychological-mechanical state into perceptible interactive signals, making it difficult for trainees to develop coordinated muscle memory and conditioned reflexes to cope with complex working conditions, and failing to meet the GWO (Global Wind Energy Organization) standards for cultivating intrinsically safe capabilities for high-risk operations. Summary of the Invention

[0005] To overcome the aforementioned shortcomings of existing technologies, this invention provides a wind turbine bolt hydraulic tightening training system and method based on the GWO standard. By collecting biological and motion data of the operator's hand to construct a virtual fluid state model, and combining enhanced visual and auditory perception with variable damping tactile feedback, a collaborative perception channel is established to overcome visual and auditory barriers. The flexible damping interlocking and bidirectional tactile interaction mechanism of this invention effectively solves the problem of coordination errors in collaborative training in confined spaces, significantly improving training safety and the trainees' collaborative understanding.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A training method for hydraulic tightening of wind turbine bolts based on the GWO standard includes:

[0008] Collect the fingertip pressure distribution matrix, head posture data, and respiratory rate of the operator; calculate the virtual fluid injection effectiveness index based on the fingertip pressure distribution matrix and head posture data; and generate the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate based on the operator's respiratory rate and the virtual fluid injection effectiveness index.

[0009] Virtual breathing light strip rendering is performed based on the virtual fluid dynamic liquid level function and the virtual fluid liquid level fluctuation rate, and background sound of fluid flow operation is generated for auditory enhancement display; the pump station operator's pressurization intention signal is detected and a visual warning is generated in combination with the virtual fluid dynamic liquid level function;

[0010] The viscosity coefficient of the virtual fluid is calculated based on the virtual fluid dynamic surface function and the virtual fluid surface fluctuation rate. The virtual fluid viscosity coefficient is converted into a damping control current value for variable damping tactile feedback control, and a two-way interactive mechanism of tactile pre-compression tightening and reverse ejection is established.

[0011] The method for calculating the virtual fluid injection effectiveness index includes:

[0012] The effective total grip pressure is obtained by performing an integral calculation on the fingertip pressure distribution matrix;

[0013] Determine whether the head posture data meets the posture conditions and whether the total effective grip pressure meets the grip conditions. When both the posture conditions and grip conditions are met, the total effective grip pressure is normalized based on the set minimum grip force threshold to obtain the virtual fluid injection effectiveness index.

[0014] When either the posture condition or the grip condition is not met, the virtual fluid injection effectiveness index is directly set to zero.

[0015] The head attitude data includes head pitch angle and head yaw angle;

[0016] The method for determining whether the attitude conditions are met is as follows: both the head pitch angle and the head yaw angle fall within the preset safe operating standard attitude space.

[0017] The method for determining whether the gripping conditions are met is: the total effective gripping pressure is greater than the minimum gripping force threshold.

[0018] The virtual fluid dynamic liquid level function uses the virtual fluid injection effectiveness index as the carrier amplitude and the breathing frequency as the modulation frequency.

[0019] The virtual fluid surface fluctuation rate is obtained by performing second derivative analysis on the virtual fluid dynamic surface function.

[0020] The method for performing virtual breathing light strip rendering includes:

[0021] Based on the length of the virtual breathing light band mapped by the virtual fluid dynamic liquid surface function, the virtual fluid liquid surface volatility is compared with a preset turbulence threshold. When the virtual fluid liquid surface volatility... The absolute value is greater than or equal to the turbulence threshold. When the virtual fluid is determined to be in a turbulent state, the surface undulation rate of the virtual fluid is... The absolute value is less than the turbulence threshold. When the virtual fluid is in a laminar flow state, the color and texture of the virtual breathing light band are determined based on whether the flow is turbulent or laminar.

[0022] The method for generating background noise during fluid flow operations includes:

[0023] The synthesized volume is controlled by the virtual fluid dynamic liquid surface function, and the synthesized timbre is controlled by the virtual fluid liquid surface fluctuation rate, thereby generating background sound for fluid flow operations.

[0024] The method for generating visual alerts by combining virtual fluid dynamic liquid level functions includes:

[0025] The manual pressurization intention signal of the pump station is used to determine the manual pressurization liquid level of the pump station. The difference between the manual pressurization liquid level of the pump station and the virtual fluid dynamic liquid level function is calculated to obtain the virtual pressure potential energy difference. The trigger condition for the visualization warning is that the virtual pressure potential energy difference is greater than zero.

[0026] The method for establishing a two-way interactive mechanism of tactile pre-compression and reverse ejection includes:

[0027] When the pump station manual pressurization controller button is detected to enter the pre-travel state, a pre-pressure tightening signal is sent to the operator training terminal, which drives the shape memory alloy component of the operator training terminal to generate a tactile tightening sensation; the effective grip total pressure is monitored in real time, and when a sudden drop in the effective grip total pressure is detected during the effective period of the pre-pressure tightening signal, the reverse ejection mechanism is triggered.

[0028] A wind turbine bolt hydraulic tightening training system based on the GWO standard is used to implement the above-mentioned wind turbine bolt hydraulic tightening training method based on the GWO standard. The system includes:

[0029] State mapping module: used to collect the fingertip pressure distribution matrix, head posture data and respiratory rate of the operator; calculate the virtual fluid injection effectiveness index based on the fingertip pressure distribution matrix and head posture data; generate the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate based on the operator's respiratory rate and the virtual fluid injection effectiveness index;

[0030] Audiovisual feedback module: used to render virtual breathing light strips based on the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate, and generate background sound for fluid flow operation for auditory enhancement display; detect the pump station operator's pressurization intention signal, and generate visual warnings in combination with the virtual fluid dynamic liquid level function;

[0031] Tactile feedback module: It is used to calculate the viscosity coefficient of virtual fluid based on the virtual fluid dynamic liquid surface function and the virtual fluid liquid surface fluctuation rate, convert the virtual fluid viscosity coefficient into the damping control current value for variable damping tactile feedback control, and establish a two-way interactive mechanism for tactile pre-compression tightening and reverse ejection.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0033] This invention constructs a virtual fluid state mapping model by collecting multidimensional physiological and behavioral data of operators. This transforms the operator's readiness state and psychological tension—difficult to accurately convey in confined spaces such as wind turbine hubs due to visual obstruction and ambient noise—into quantifiable and transferable virtual fluid dynamic surface functions and surface fluctuation rates. Furthermore, these fluid characteristic parameters are mapped in real-time to the variable damping tactile feedback and audiovisual enhancement signals of the pump station operator's pressurization controller. This solution establishes a bio-mechanical coupling perception channel between physically isolated operators, independent of traditional voice communication. By dynamically adjusting the mechanical damping of control buttons using the virtual fluid viscosity coefficient, the pump station operator can perceive the mechanical damping through their fingers. The "lightness" or "heaviness" of the pressure directly senses the stability of the operator's grip and the smoothness of their breathing, realizing a shift in the interaction paradigm from relying on subjective "voice confirmation" to objective "tactile interlocking." At the same time, the two-way interactive mechanism of tactile pre-pressure tightening and reverse ejection based on state data can, within milliseconds, gently dissuade or rigidly brake the pump station operator's intention to pressurize by increasing physical pressing resistance or generating a reverse impact force when the operator has not reached the optimal working state or has abnormally released their grip. This effectively eliminates the coordination time difference under visual and auditory blind spots without the need for verbal communication, significantly improving the inherent safety level and the accuracy of action execution in high-risk confined space two-person collaborative operations. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 A flowchart illustrating the principle of the wind turbine bolt hydraulic tightening training method based on the GWO standard provided in this embodiment of the invention;

[0036] Figure 2 This is a schematic diagram of a confined space operation scenario in a wind turbine hub, provided by an embodiment of the present invention.

[0037] Figure 3 This is a schematic diagram of the sensor distribution on the training gloves provided in an embodiment of the present invention;

[0038] Figure 4 A flowchart illustrating the principle of triggering visual alerts provided in this embodiment of the invention;

[0039] Figure 5A functional block diagram of a wind power bolt hydraulic fastening training system based on the GWO standard provided in an embodiment of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Example 1

[0042] Please see Figure 1 As shown, this embodiment provides a training method for hydraulic tightening of wind turbine bolts based on the GWO standard, including:

[0043] Step S10: Collect the fingertip pressure distribution matrix, head posture data, and respiratory rate of the operator; calculate the virtual fluid injection effectiveness index based on the fingertip pressure distribution matrix and head posture data; generate the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate based on the operator's respiratory rate and the virtual fluid injection effectiveness index.

[0044] Specifically, step S10 constructs a virtual fluid state mapping model based on the coupling of biometrics and mechanical actions. Its core lies in abstracting the physiological state and mechanical behavior of the operator performing hydraulic bolt tightening operations within the confined space of the wind turbine hub into a dynamic process of injecting fluid into a virtual communicating vessel. Please refer to [link to relevant documentation]. Figure 2 The diagram shown is a schematic of a confined space operation scenario for a wind turbine hub provided in this embodiment. The diagram illustrates the spatial distribution structure inside and outside the hub. As the connection point between the wind turbine blades and the main shaft, the wind turbine hub typically has a closed internal space, the diameter of which varies depending on the turbine capacity. Operators must perform hydraulic tightening of high-strength bolts within this confined space. In such operation scenarios, such as... Figure 2 As shown, the operator responsible for holding the hydraulic wrench is located inside the wheel hub, while the pump station operator responsible for controlling the hydraulic pump station is located outside the wheel hub. There is physical isolation between the two. The operator, located inside the wheel hub, performs wrench positioning and gripping operations, while the pump station operator, located outside the wheel hub or on the tower platform, controls the pressurization and pressure-holding actions of the hydraulic system. Combined with... Figure 2 As indicated by the visual barrier markings between the operator and the pump station operator, the view between the two is completely blocked by the hub housing. Furthermore, the mechanical noise generated during the operation of the hydraulic pump station, combined with the resonance effect inside the hub, causes severe distortion of the voice communication signal, making it difficult for traditional walkie-talkies to accurately convey the operator's readiness status information.

[0045] The virtual fluid state mapping model draws inspiration from the communicating vessels principle in classical fluid mechanics, analogizing the operator's effective gripping action to the process of injecting fluid into one end of a communicating vessel. The greater the gripping force and the more stable the posture, the more virtual fluid is injected, causing the liquid level inside the vessel to rise. This analogy transforms the difficult-to-perceive muscle strength and postural stability into quantifiable and transferable liquid level information, allowing the pump operator to indirectly perceive their real-time readiness by observing the virtual liquid surface. Respiratory rate, as an external manifestation of the human autonomic nervous system, exhibits periodic fluctuations superimposed on the virtual liquid surface, creating dynamic undulations. When breathing is steady, the liquid surface undulations are gentle, exhibiting a laminar flow state; when breathing is rapid or the hand trembles slightly, the liquid surface undulates violently, exhibiting a turbulent flow state, thus visualizing the operator's physiological tension.

[0046] Further, step S10 includes:

[0047] Step S11: Collect the fingertip pressure distribution matrix of the operator's hand through the training glove with integrated distributed pressure sensor array and calculate the effective grip total pressure by integration. Obtain head posture data including head pitch angle and head yaw angle through the safety helmet with posture capture function. Calculate the virtual fluid injection effectiveness index based on the effective grip total pressure, head pitch angle and head yaw angle.

[0048] Specifically, the training glove with an integrated distributed pressure sensor array employs flexible thin-film pressure sensor technology. (See also...) Figure 3 This is a schematic diagram of the sensor distribution on the training glove provided in this embodiment. Figure 3 The diagram illustrates the fingertip sensors located at the tips of the five fingers and the palm sensor area in the middle of the palm, with dots indicating pressure sampling points. The sensor array is deployed in the fingertip areas and palm area of ​​the glove, forming a... Figure 3 The diagram shows a pressure acquisition grid covering the main gripping contact surface. The fingertip pressure distribution matrix represents the real-time pressure readings of each sensor node in two-dimensional matrix form; the number of rows and columns of the matrix is ​​determined by the physical layout of the sensor array. For example, combined with... Figure 3 The distribution shown indicates that if the sensor array has four sampling points on each finger, there will be a total of twenty sampling points across the five fingers. Figure 3The fingertip pressure distribution matrix formed by several sampling points in the palm area can be represented as an M×N dimensional matrix, where M is the number of sensor rows and N is the number of sensor columns. The effective grip total pressure is calculated by integrating the fingertip pressure distribution matrix, weighting the pressure readings of all sensor nodes in the matrix according to their corresponding effective sensor areas. The physical significance of integration is to transform discrete local pressure values ​​into a single value representing the overall grip strength, facilitating subsequent comparison with thresholds. The integration formula can be expressed as the effective grip total pressure equals the sum of the products of the pressure values ​​of each sensor node and their corresponding effective areas; that is, multiplying each element in the fingertip pressure distribution matrix by its corresponding area element and then summing the results. The helmet with attitude capture function has a built-in inertial measurement unit (IMU), composed of a three-axis accelerometer and a three-axis gyroscope, which can calculate the helmet's spatial attitude relative to the direction of gravity in real time. The head pitch angle is defined as the angle at which the helmet tilts forward or backward; a positive value indicates a forward tilt and a negative value indicates a backward tilt. The head yaw angle is defined as the angle at which the helmet turns left or right; a positive value indicates a right turn and a negative value indicates a left turn. The attitude calculation of the inertial measurement unit uses a quaternion algorithm or a complementary filtering algorithm to fuse the raw data from the accelerometer and gyroscope, eliminating the cumulative drift error of a single sensor and outputting stable angle information.

[0049] The safe operating standard posture space is a predefined permissible range of head posture, with its boundary values ​​determined according to the safe operating procedures for bolt tightening in the GWO Basic Skills Training Standard. When operating a hydraulic wrench inside a wheel hub, the operator needs to visually monitor the engagement position of the wrench and bolt to ensure the socket is fully engaged with the nut. The head pitch and yaw angles should be maintained within a range that allows for clear observation of the workpiece. For example, the safe operating standard posture space can be set as a head pitch angle between -30° and +45° and a head yaw angle between -45° and +45°. Exceeding this range indicates that the operator's line of sight has deviated from the work area, possibly due to posture adjustments or distraction.

[0050] The calculation of the virtual fluid injection effectiveness index combines conditional judgment and normalization. The judgment conditions include two aspects: attitude conditions require that the real-time acquired head pitch and yaw angles fall within the safe operating standard attitude space; and grip conditions require that the total effective grip pressure is greater than a preset minimum grip force threshold. The minimum grip force threshold is determined based on the minimum safe grip requirements for hydraulic wrench operation. When the grip force is below this threshold, the wrench may slip during hydraulic pressurization, causing personal injury or equipment damage. For example, the minimum grip force threshold can be set as the minimum grip strength that can effectively resist the reaction torque of the hydraulic wrench. When both attitude and grip conditions are met, the total effective grip pressure is normalized based on the minimum grip force threshold to obtain the virtual fluid injection effectiveness index. The normalization method involves subtracting the minimum grip force threshold from the total effective grip pressure and then dividing by the grip force range, thus limiting the value range of the virtual fluid injection effectiveness index to between 0 and 1. The gripping force range is defined as the difference between the maximum gripping force Pmax that the pressure sensor array can detect and the minimum gripping force threshold Pmin, i.e., the gripping force range equals Pmax - Pmin. When any condition is not met, the virtual fluid injection effectiveness index is directly set to zero, indicating that the operator is not in an effective working state at the current moment, which is equivalent to the virtual fluid injection channel being closed and the liquid level in the communicating vessel not rising.

[0051] Step S11 integrates the operator's mechanical motion characteristics and physiological posture characteristics into a single quantitative index. A conditional judgment mechanism ensures that a valid state signal is generated only when the operator is in a safe posture and the grip strength meets requirements. This design allows subsequent steps to determine whether the operator has completed preparation based on the virtual fluid injection effectiveness index, avoiding the information distortion problem caused by relying on subjective voice confirmation in traditional solutions. The dual constraint of the conditional judgment binds posture safety and grip effectiveness, so that a virtual fluid injection effectiveness index of zero not only indicates insufficient grip strength but may also indicate posture deviation, thus filtering out false trigger signals under unsafe conditions at the source. Without the data acquisition and conditional judgment mechanism of step S11, the subsequent virtual fluid dynamic liquid level function would lack accurate amplitude input, potentially leading the pump operator to misjudge the operator's readiness before the operator has completed grip preparation, thereby triggering pressurization and creating a safety hazard.

[0052] Step S12: Use a respiratory monitoring device to obtain the operator's breathing frequency, use the virtual fluid injection effectiveness index as the carrier amplitude and the breathing frequency as the modulation frequency to construct a virtual fluid dynamic liquid surface function that characterizes the liquid level change in the virtual communicating vessel, and perform second derivative analysis on the virtual fluid dynamic liquid surface function to calculate the virtual fluid liquid surface volatility.

[0053] Specifically, the respiratory monitoring device uses a chest-abdomen breathing sensor or an airflow sensor integrated into the chin strap of a helmet to collect respiratory signals. The chest-abdomen breathing sensor uses elastic strain gauges to detect the periodic fluctuations of the chest or abdominal wall, while the airflow sensor identifies the respiratory cycle by detecting temperature or flow rate changes in exhaled air from the mouth and nose. Respiratory rate extraction employs zero-crossing detection or peak detection methods, performing periodic analysis on the respiratory signal waveform and calculating the reciprocal of adjacent respiratory cycles to obtain the instantaneous respiratory rate. For example, the respiratory rate of an adult at rest is typically between twelve and twenty breaths per minute, equivalent to approximately 0.2 to 0.33 Hz; during physical labor or periods of mental stress, the respiratory rate can increase to twenty-four to thirty-six breaths per minute, equivalent to approximately 0.4 to 0.6 Hz.

[0054] The virtual fluid dynamic surface function is constructed using an amplitude modulation mathematical model. The virtual fluid injection effectiveness index calculated in step S11 is used as the carrier amplitude, and the real-time breathing frequency is used as the modulation frequency. The mathematical expression for the virtual fluid dynamic surface function is: In the formula, L(t) represents the virtual fluid dynamic liquid level function value at time t, characterizing the height of the liquid level in the virtual communicating vessel relative to the reference liquid level. Its value range is constrained by both the virtual fluid injection effectiveness index and the breathing depth coefficient A. E inject The virtual fluid injection effectiveness index, ranging from 0 to 1, characterizes the combined effectiveness of the operator's current grip and posture. This value serves as the base height of the liquid surface, determining the overall liquid level. A is the breathing depth coefficient, representing the weight of breathing fluctuations on the amplitude of liquid surface fluctuations. Its value is determined based on the correlation between the amplitude of chest and abdominal movements during human respiration and grip stability. For example, when A is set to 0.15, the amplitude of liquid surface fluctuations caused by breathing is 15% of the base liquid surface height. This proportion makes the visualization of breathing rhythm sufficiently significant without obscuring the dominant information about grip force changes. breath The real-time respiratory rate, measured in Hertz, is collected and updated in real time by a respiratory monitoring device. This parameter determines the rate of periodic fluctuations in the liquid level; the frequency of liquid level fluctuations increases with faster breathing. t is a time variable, measured in seconds, used to calculate the phase of the sine wave. It is a phase constant used to calibrate the initial phase relationship between the respiratory signal and the liquid level fluctuation. It is usually calibrated during system initialization based on the first complete respiratory cycle.

[0055] The formula's design logic lies in decomposing the operator's ready state into a superposition of static and dynamic components. The static component is represented by the virtual fluid injection effectiveness index E. injectThe static component, reflecting the stability of the operator's grip and posture, changes relatively slowly, only significantly altering when the operator adjusts their grip or posture. The dynamic component, generated by the breathing modulation term, is a sine wave superimposed on the static component, reflecting the operator's physiological rhythm and mental state. When the operator's breathing is stable, the sine wave has a longer period and uniform amplitude, resulting in a regular, slow fluctuation in the liquid surface. When the operator's breathing is rapid or irregular, the sine wave's period shortens and there are differences between adjacent periods, resulting in rapid and irregular fluctuations in the liquid surface. The expression within square brackets ensures that the dynamic component fluctuates symmetrically around the static component, with the highest point of the liquid surface being E. inject ×(1+A), the lowest point of the liquid surface is E. inject ×(1-A) ensures that the liquid level is always non-negative and the range of change is controllable.

[0056] The method for calculating the volatility of the virtual fluid surface is as follows: Second derivative analysis is performed on the dynamic surface function of the virtual fluid. Mathematically, the second derivative characterizes the concavity and convexity of the function curve and the rate of curvature change. For periodic oscillation functions, the absolute value of the second derivative reflects the acceleration characteristics of the oscillation. The physical meaning of the virtual fluid surface volatility lies in quantifying the severity of surface changes. High volatility indicates that the surface is undergoing rapid acceleration or deceleration, analogous to turbulent disturbances in the fluid; low volatility indicates gentle surface changes, analogous to laminar flow. The analytical result of obtaining the second derivative of the dynamic surface function of the virtual fluid is as follows:

[0057] ;

[0058] in, The above formula shows that the amplitude of the virtual fluid surface volatility is related to the virtual fluid injection effectiveness index E. inject Respiratory depth coefficient A, respiratory rate f breath It is proportional to the square of . It is a sine function. When the operator's hand is holding the grip effectively and breathing is rapid, E inject Larger and f breath The amplitude of the virtual fluid surface fluctuation rate is also relatively large, leading to a significant increase in the amplitude of the fluctuation rate, indicating that although the operator has completed the grip, the physiological state is tense, and it is not the optimal time to apply pressure. When the operator's grip is effective and breathing is calm, E inject t is larger and f breathThe virtual fluid surface volatility amplitude remains low, indicating that the operator is in a stable state and suitable for hydraulic pressurization. Besides the regular fluctuations caused by sinusoidal modulation, the operator's hand tremors also introduce surface volatility components. The frequency of these hand tremors is higher than the normal breathing frequency, manifesting as high-frequency disturbances superimposed on the low-frequency envelope of the breathing modulation. The virtual fluid surface volatility amplifies high-frequency components because the frequency is squared in the amplitude calculation during second-order derivative operations; therefore, the volatility increment caused by high-frequency tremors is significantly greater than its representation in the first derivative or the original signal. This characteristic allows the virtual fluid surface volatility to sensitively detect operator muscle fatigue or mental stress. Even if the grip strength remains above a threshold, a spike in volatility caused by hand tremors can still alert the pump operator to delay pressurization operations.

[0059] Step S12 fuses the static gripping state information output from step S11 with the dynamic physiological rhythm information acquired from respiratory monitoring to generate a composite state function that simultaneously includes the degree of readiness and the degree of stability. The amplitude of the virtual fluid dynamic surface function reflects whether the operator has completed gripping preparation, and the fluctuation characteristics reflect whether the operator is in a physiologically stable state. The combination of the two allows the pump station operator to determine that the operator not only grips the wrench but also grips it firmly and maintains a stable mindset. The virtual fluid surface fluctuation rate is extracted as an independent indicator to provide turbulence criteria for visual rendering and auditory synthesis in the subsequent step S20, enabling the light band color and background sound effects to intuitively reflect the degree of stability, rather than just the gripping force. The breathing frequency is incorporated as a modulation parameter into the virtual fluid dynamic surface function. In step S22, the auditory entrainment effect can be used to induce the pump station operator to unconsciously adjust their breathing rhythm to match the operator, thereby establishing synchronicity between the two at the physiological level. This is an implicit coordination mechanism that traditional voice communication schemes cannot achieve. If the breathing modulation and fluctuation rate calculation in step S12 is missing, the status information output by the system only contains the static value of the grip force, and cannot distinguish whether the operator is holding calmly or holding tensely. The pump station operator may misjudge that the operator is ready when the operator is holding his breath and performing pressurization, resulting in a dangerous situation where the operator releases his grip due to the impact of the reaction force.

[0060] Step S13: Package the virtual fluid dynamic liquid level function, the virtual fluid liquid level fluctuation rate, and the effective grip total pressure to generate a collaborative state data packet, and synchronously distribute the collaborative state data packet to the operator training terminal and the pump station operator training terminal.

[0061] Specifically, the collaborative state data packet encapsulates the multi-dimensional state information generated in steps S11 and S12 into a unified data unit according to a predefined data structure, facilitating efficient transmission and parsing in a distributed training system. The data packet's field structure includes the current sampled value of the virtual fluid dynamic surface function L(t), the virtual fluid surface volatility, and other parameters. The current calculated value, effective holding total pressure P total The system timestamp of the sampling time and the data packet sequence number are used to detect data packet loss and trigger a retransmission mechanism. Data packet transmission relies on the industrial Ethernet architecture of the GWO basic skills training platform. Industrial Ethernet uses the real-time Ethernet protocol, which, compared to standard Ethernet, features deterministic transmission delay and high reliability, making it suitable for latency-sensitive control and feedback applications. The operator training terminal is configured as the master node, acting as the data packet generation node, and the pump station operator training terminal is configured as the slave node, acting as the data packet receiving and processing node. At the end of each sampling period, the master node broadcasts or directs the encapsulated collaborative status data packet to the slave node via the industrial Ethernet. After receiving the data packet, the slave node parses the information in each field for local visualization rendering and haptic feedback control. In addition to being sent to the pump station operator training terminal, the collaborative status data packet is also simultaneously sent to the operator training terminal for local display and status confirmation. After receiving the data packet, the operator training terminal can simultaneously render a virtual breathing light strip on the AR display device, allowing the operator to observe a visual representation of their own status, forming a visual feedback loop. This design allows the operator to autonomously adjust their grip strength and breathing rhythm based on the length and color of the light strip, proactively coordinating with the pump operator at a conscious level, rather than passively waiting for the pump operator's pressurization operation. Step S13 completes the transformation from single-machine state acquisition to distributed state sharing, synchronously transmitting the biomechanical coupling state information generated on the operator's side to the pump operator's side, breaking down the information barrier between the two due to visual isolation and noise interference. The adoption of industrial Ethernet provides data transmission with industrial-grade reliability and deterministic latency characteristics, avoiding signal interference and transmission jitter problems that may exist in wireless communication. Without the data packaging and distribution mechanism of step S13, the state information generated in steps S11 and S12 would only exist locally on the operator's training terminal, and the pump operator would not be able to obtain any state information from the operator. The subsequent visualization rendering in step S20 and the tactile feedback in step S30 would lose their data source, and the entire collaborative training scheme based on state sharing would not be feasible.

[0062] Step S10 forms a complete state perception and transmission chain from data acquisition, feature extraction, model construction to information distribution, transforming the operator's preparation state, which is difficult to directly observe and describe in a confined space, into quantifiable, transmittable, and perceptible virtual fluid state information. The acquisition and integral calculation of the fingertip pressure distribution matrix aggregates distributed grip force information into a single effective total grip pressure index, eliminating the interference of uneven pressure distribution caused by differences in grip posture on state judgment, allowing subsequent calculations to be processed based on a unified grip force standard. The introduction of head posture data embeds safety constraints into the state evaluation logic, ensuring that the virtual fluid injection effectiveness index only generates a valid value when the operator's posture conforms to safety specifications, eliminating the possibility of false triggering under posture deviation, and establishing the entire collaborative mechanism on the premise of safe operation. The modulation and fusion of respiratory rate and the virtual fluid injection effectiveness index superimposes static grip preparation information and dynamic physiological rhythm information into the same function, allowing the virtual fluid dynamic liquid surface function to simultaneously carry information on both whether the operator is ready and whether the operator is stable. The second derivative analysis of the virtual fluid surface turbulence amplifies high-frequency disturbances and can sensitively detect stability decreases caused by hand tremors and rapid breathing, providing a quantitative criterion for distinguishing between laminar and turbulent flow states in subsequent steps. This design allows the pump operator to not only determine if the operator has gripped the wrench, but also whether the grip is stable and whether the operator is calm, thus selecting to perform pressurization when the operator is in the optimal state, rather than relying solely on a simple binary judgment based on whether the grip strength meets the standard.

[0063] The generation and distribution of collaborative status data packets transforms single-point acquired status information into a distributed, shared collaborative foundation. The industrial Ethernet transmission architecture and time synchronization mechanism ensure that the master and slave nodes maintain consistent perception of status changes. This information-sharing mechanism enables operators and pump station operators to make their respective reactions and decisions based on the same status data, forming a collaborative condition with symmetrical information. The data packet simultaneously retains three types of information: the virtual fluid dynamic surface function, the virtual fluid surface fluctuation rate, and the effective grip total pressure. This provides multi-level data input for the subsequent audiovisual rendering in step S20 and the tactile feedback in step S30, allowing different sensing channels to select appropriate status representations for presentation and control based on their respective characteristics.

[0064] Step S10 transforms the abstract physiological state of the human body and the mechanical operation state into a concrete metaphor of fluid dynamics. The rise and fall of the liquid level in the virtual communicating vessel and the laminar and turbulent states of the fluid can be intuitively understood. Operators without professional knowledge can grasp the rhythm of coordination by observing the liquid level. This metaphorical design reduces the cognitive load of collaborative training, allowing trainees to focus on the task itself rather than complex data interpretation, and establish a physiological tacit understanding for collaboration between the two at a subconscious level. The introduction of the breathing modulation mechanism gives the virtual fluid dynamic liquid level function a natural periodic fluctuation. This fluctuation is synchronized with the human breathing rhythm. In the auditory enhancement display in step S22, it can induce the pump operator to unconsciously adjust their breathing to match the operator's rhythm, thereby achieving cross-space rhythm synchronization at the physiological level. This is a deep collaborative effect that traditional voice commands or signal light schemes cannot achieve.

[0065] Step S20: Render a virtual breathing light strip based on the virtual fluid dynamic surface function and the virtual fluid surface fluctuation rate, and generate background sound for fluid flow operation for auditory enhancement display; detect the pump station operator's pressurization intention signal and generate a visual warning based on the virtual fluid dynamic surface function;

[0066] Step S20 converts the virtual fluid dynamic surface function and virtual fluid surface volatility contained in the collaborative state data packet into visual and auditory signals that can be synchronously perceived by both the operator training terminal and the pump station operator training terminal. When performing hydraulic bolt tightening operations within the confined space of the wind turbine hub, there is physical isolation between the operator and the pump station operator. The pump station operator cannot directly observe whether the operator has completed wrench positioning, whether the grip strength is sufficient, or whether the physiological state is stable. Traditional walkie-talkie communication is also unable to accurately transmit the operator's readiness status due to severe distortion caused by environmental noise. Step S20 establishes a non-verbal perception channel that bypasses voice communication by constructing a visual rendering of a virtual breathing light strip and auditory synthesis of the background sound of fluid flow operations. This allows the pump station operator to judge the operator's grip readiness level by changes in the length of the light strip, judge the operator's physiological stability by changes in the color and texture of the light strip, and perceive the operator's breathing rhythm and tension level by changes in the timbre of the background sound effects. This audiovisual fusion state presentation method transforms the abstract numerical information in step S10 into intuitively perceptible environmental signals, enabling pump station operators to subconsciously grasp the real-time state of the operator without having to deliberately interpret the data, and thus choose to perform the pressurization operation when the operator is in a stable and ready state.

[0067] Further, step S20 includes:

[0068] Step S21: Extract the virtual fluid dynamic surface function and virtual fluid surface turbulence from the cooperative state data packet; map the length of the virtual breathing light band according to the virtual fluid dynamic surface function; compare the virtual fluid surface turbulence with the preset turbulence threshold; and determine the color texture of the virtual breathing light band according to the comparison result.

[0069] Specifically, the virtual breathing light strip is a dynamic visual element synchronously rendered at the edge of the field of view of the AR display devices equipped on the operator training terminal and the pump station operator training terminal. Its position is set in the peripheral area of ​​the field of view rather than the center. This positioning is based on the fact that the peripheral visual area is highly sensitive to movement and change. Trainees can perceive the changes in the light strip's state through their peripheral vision without taking their eyes off the work object, avoiding distraction from the work site caused by focusing on the light strip. The AR display device can be a head-mounted augmented reality glasses or a transparent display unit integrated into the safety helmet goggles. The displayed content is superimposed on the real work scene, allowing trainees to simultaneously perceive the changes in the virtual breathing light strip while observing the engagement status of bolts and wrenches.

[0070] The length mapping of the virtual breathing light strip adopts a linear proportional relationship, mapping the real-time value of the virtual fluid dynamic liquid level function L(t) in the cooperative state data packet to the extension length of the light strip in the display field of view. The mapping formula is: Light strip length L bar =L(t) × maximum length of light band L bar,max The maximum length L of the light band bar,max The value range of the virtual fluid dynamic liquid level function L(t) is determined by the visible area size and the position of the light strip of the AR display device. This is because the range of values ​​for the virtual fluid injection effectiveness index E is affected by the virtual fluid injection effectiveness index E. inject Constrained by the breathing depth coefficient A, when E inject When full scale is reached and peak inspiration is achieved, L(t) reaches E. inject The upper limit of ×(1+A) is when the light band extends to its maximum length; when E inject When L(t) is zero or at the trough of exhalation, it drops to its lower limit, and the light band contracts to its shortest length or even disappears. This length mapping synchronizes the expansion and contraction of the light band with the operator's grip strength and respiratory fluctuations. When the operator inhales and tightens the wrench, the light band extends; when the operator exhales and relaxes the grip, the light band contracts, creating a concrete simulation of the process of injecting virtual fluid into the communicating vessel. The pump operator can perceive the operator's respiratory cycle by observing the expansion and contraction rhythm of the light band. When the light band extends to a higher position and remains stable, it is determined that the operator has completed gripping preparation and is in an inhalation and energy storage state, which is the appropriate time to perform the pressurization operation.

[0071] The color and texture rendering of the virtual breathing light band is based on the virtual fluid surface fluctuation rate. The result is determined by comparing it with a preset turbulence threshold. Turbulence threshold The determination is based on the statistical distribution characteristics of the virtual fluid surface volatility. During the system calibration phase, virtual fluid surface volatility data from multiple trainees under different operating conditions are collected. The differences in volatility distribution between calm and tense gripping states are analyzed, and the critical value that can effectively distinguish between the two states is set as the turbulence threshold. For example, the upper quartile of the virtual fluid surface volatility distribution under calm gripping state can be used as the turbulence threshold, so that about 3 / 4 of the calm state samples fall below the turbulence threshold and about 3 / 4 of the tense state samples fall above the turbulence threshold, thereby establishing a clear distinction boundary between the two states.

[0072] Color and texture rendering uses conditional branching logic: when the virtual fluid surface undulation rate... The absolute value is greater than or equal to the turbulence threshold. When the system determines that the virtual fluid is in a turbulent state, it renders the virtual breathing light band as an amber-colored noise texture. Amber belongs to the warning color family of warm colors, which has a psychological effect of attracting attention without causing panic in human visual perception. The noise texture uses randomly distributed bright and dark spots to simulate the visual effect of suspended particles in turbid liquid. The flickering frequency of the spots is proportional to the fluctuation rate of the virtual fluid surface; the greater the fluctuation rate, the more intense the spot flickering, visually demonstrating the turbulent state of the fluid. When the fluctuation rate of the virtual fluid surface... The absolute value is less than the turbulence threshold. When the system determines that the virtual fluid is in a laminar flow state, it further determines whether the virtual fluid dynamic liquid level function L(t) is maintained at a high level. The high level determination condition is that L(t) is greater than a preset proportion of its full-scale value and the duration exceeds a preset maintenance duration. The full-scale value is defined as the virtual fluid injection effectiveness index E. inject The value of L(t) is set at full scale and when breathing is at its peak inhalation. The preset ratio and preset maintenance duration are determined based on the time requirements for bolt tightening operation preparation confirmation in the GWO basic skills training standards. For example, the preset ratio can be set to 80% of the full scale value, and the preset maintenance duration can be set between 1 and 2 seconds. When the above high-level maintenance conditions are met, the virtual breathing light band is gradually rendered into a high-saturation emerald green smooth fluid texture. Emerald green has a psychological suggestion of safety and passage in human visual perception. The smooth fluid texture uses a continuous gradient of light and dark transitions to simulate the smooth flow effect of clear liquid, indicating that the operator has completed the grip preparation and the physiological state is stable, and the pump station operator can safely perform the pressurization operation. The gradient rendering method, rather than the instantaneous switching method, makes the color change process smooth and natural, avoiding the visual shock and misjudgment caused by abrupt changes.

[0073] Step S21 transforms the numerical state information generated in step S10 into visually perceptible light band shape and color information. Length mapping allows the pump station operator to intuitively judge the hand's grip readiness and breathing rhythm, while color texture mapping allows the operator to intuitively judge the operator's physiological stability. This visual presentation method has a richer information hierarchy compared to traditional traffic light-style binary signals. The continuous change in light band length carries fine information about grip strength and breathing phase, while the laminar and turbulent distinctions in color texture carry stability judgment information. The combination of these two allows the pump station operator to perceive that the operator not only grips the wrench but also holds it firmly and breathes smoothly. Without the visual rendering mechanism of step S21, the pump station operator would not be able to intuitively perceive the state data generated in step S10. Although the tactile feedback in subsequent step S30 can provide tactile information about damping changes, the information bandwidth of the tactile channel is limited and requires the pump station operator to actively try pressing the button to perceive it. The visual information provided by step S21 allows the pump station operator to predict the operator's state before even touching the button, forming a dual guarantee of visual prediction and tactile confirmation.

[0074] Step S22: The synthesized volume is controlled by the virtual fluid dynamic liquid surface function, and the synthesized timbre is controlled by the virtual fluid liquid surface fluctuation rate to generate the background sound of fluid flow operation. The background sound of fluid flow operation is superimposed on the communication headset for auditory enhancement display.

[0075] Specifically, the background noise of fluid flow operations is an audio signal synthesized in real time by the system's built-in fluid acoustic engine. Its volume and timbre parameters are dynamically controlled by the virtual fluid dynamic surface function and virtual fluid surface fluctuation rate in the collaborative state data packet. The fluid acoustic engine employs physical modeling synthesis technology, incorporating a built-in pipe fluid acoustic model, and can generate simulated liquid flow sound effects in pipes in real time based on input parameters. Volume control uses the virtual fluid dynamic surface function L(t) as the gain coefficient to synthesize the volume. , where V base This represents the reference volume. The reference volume is determined based on the audible threshold and comfortable loudness range of the communication headset in the background noise of a wind turbine hub operating environment. The lowest volume value at which trainees can clearly distinguish the background sound of fluid flow in a typical operating noise environment is determined through human factors engineering experiments and used as the audible threshold. The highest volume value at which trainees report no auditory fatigue is used as the comfortable upper limit. The reference volume is set as the geometric mean of the audible threshold and the comfortable upper limit; L max The full-scale value of the virtual fluid dynamic liquid surface function L(t) is represented. When L(t) is large, the synthesized volume is high, and when L(t) is small, the synthesized volume is low, so that the loudness change of the background sound is synchronized with the change of the light band length, forming a consistent audiovisual feedback.

[0076] The timbre control uses virtual fluid surface fluctuation rate. With turbulence threshold The comparison results are used as the criterion for timbre switching. When the virtual fluid surface fluctuation rate... The absolute value is greater than the turbulence threshold. At this time, the fluid acoustic engine synthesizes and plays turbulent sound effects. The acoustic characteristics of the turbulent sound effects include two components: bubble bursting sound and pipe friction sound. The bubble bursting sound uses a random pulse generator to generate irregular, short popping sounds. The frequency of the popping sounds is proportional to the amplitude of the virtual fluid surface fluctuation rate; the greater the fluctuation rate, the more frequent the popping sounds. The pipe friction sound uses bandpass filtered white noise to simulate the turbulent friction between the fluid and the pipe wall. The overall turbulent sound effects present a chaotic and irregular auditory characteristic, psychologically suggesting an unstable connection and an unprepared state, prompting the pump operator to instinctively delay operation. When the virtual fluid surface fluctuation rate... The absolute value is less than or equal to the turbulence threshold. At that time, the fluid acoustic engine synthesizes and plays laminar humming sound. The acoustic characteristics of laminar humming sound are low-frequency humming with a single frequency and rich harmonics. The fundamental frequency is the breathing frequency f in the breathing modulation component of the virtual fluid dynamic liquid surface function L(t). breath The decision was made to synchronize the periodic fluctuations of the humming sound with the operator's breathing rhythm. The harmonic structure adopts a spectrum distribution similar to the steady current sound of a high-power transformer, presenting a stable, regular, and predictable auditory characteristic. On a psychological level, this implies that the channel is clear and the state is ready, prompting the pump station operator to have an intuitive reaction that the operation is possible.

[0077] Background noise from fluid flow operations is superimposed onto a communication headset, mixing with any walkie-talkie conversations that may be present in the headset. In the wind turbine hub operating environment, the high-intensity mechanical noise generated by the hydraulic pump station is transmitted to the operator's ears through the hub shell. Traditional walkie-talkie voice signals are severely distorted in this noise environment, making it difficult to accurately convey semantic information. The spectral characteristics of the background noise from fluid flow operations differ from those of the mechanical noise. The low-frequency harmonic components of the laminar flow hum complement the high-frequency mechanical noise, allowing trainees to clearly distinguish the timbre changes of the background noise in a noisy environment. The continuous playback of the background noise creates an auditory masking effect, partially blocking the interference of environmental mechanical noise on the trainees and reducing the psychological stress caused by the noisy environment.

[0078] The design, which synchronizes the periodic fluctuations of the laminar flow hum with the operator's breathing rhythm, utilizes the physiological mechanism of the auditory entrainment effect. The auditory entrainment effect refers to the phenomenon where the human body's physiological rhythm unconsciously tends to synchronize with external periodic sound stimuli. When the pump station operator continuously receives the laminar flow hum synchronized with the operator's breathing rhythm, the pump station operator's breathing rhythm gradually adjusts subconsciously to match the periodicity of the hum, thus indirectly achieving synchronization with the operator's breathing. When the breathing rhythms of both individuals become consistent, their physiological and psychological states also tend to be coordinated. After perceiving the laminar flow hum and unconsciously adjusting their breathing, the pump station operator also enters a relatively stable physiological state. At this time, the hand movements performing the pressurization operation become more stable and precise, forming a positive cycle where the operator stably guides the pump station operator to remain stable.

[0079] Step S22 transforms the state data generated in step S10 into auditory-perceptible background sound effects, forming a multi-channel state presentation with audiovisual fusion together with the visual rendering in step S21. The parallel presentation of the visual and auditory channels has a dual function of information redundancy and complementarity: when the trainee cannot directly look at the AR display device due to work posture limitations, the auditory channel can still continuously provide state information; when environmental noise causes partial distortion of auditory information, the visual channel can still provide clear state feedback. The consistent changes in the two channels enhance the credibility and perceptual intensity of the state information, making the pump station operator more certain of the operator's state. Without the auditory synthesis mechanism in step S22, step S20 would only rely on the visual channel to provide state feedback. The visual channel requires the trainee to actively observe to obtain information, while the information reception of the auditory channel is passive and continuous. Step S22 enables the state information to be continuously transmitted through background sound effects even when the trainee's attention is focused on the work object, forming full-time state perception coverage.

[0080] Step S23, see Figure 4 The system detects the intention signal of manual pressurization at the pump station, calculates the virtual pressure potential energy difference by combining the virtual fluid dynamic liquid level function in the collaborative status data packet, and triggers a visual warning based on the virtual pressure potential energy difference.

[0081] The calculation method for virtual pressure potential energy difference includes: determining the manual pressurization liquid level of the pump station based on the manual pressurization intention signal, and calculating the difference between the manual pressurization liquid level of the pump station and the virtual fluid dynamic liquid surface function to obtain the virtual pressure potential energy difference.

[0082] Specifically, the detection of the pump operator's pressurization intention signal relies on a capacitive touch sensor integrated into the button surface of the pump operator's pressurization controller in the pump operator training terminal. The capacitive touch sensor uses the mutual capacitance detection principle, forming a capacitance detection area on the button surface. When a human finger approaches or touches the button surface, the parasitic capacitance formed between the finger and the detection electrode changes the capacitance value of the detection circuit. The sensor determines the presence of a finger by detecting this capacitance change. The capacitive touch sensor and the button's mechanical travel detection are independent. Mechanical travel detection uses a displacement sensor or microswitch to detect whether the button is pressed, while the capacitive touch sensor detects whether a finger is touching the button surface regardless of whether the button is pressed. The logic for generating the pump operator's pressurization intention signal is as follows: when the capacitive touch sensor detects a finger touching the button surface and the mechanical travel detection shows that the button is not pressed, the system determines that the pump operator has placed their finger on the button surface but has not yet performed a pressing action. At this time, a pump operator pressurization intention signal is generated, indicating that the pump operator has the intention to press but has not yet performed the pressurization operation. This pre-contact detection mechanism can capture the pump operator's pressurization intention before they actually press the button, providing a trigger condition for subsequent visual warnings.

[0083] Virtual pressure potential energy difference The calculation is based on the physical metaphor of the liquid level difference in communicating vessels. The pump station manual pressurization level Lpump is defined as a dummy variable characterizing the intensity of the pump station manual pressurization intention: when no pump station manual pressurization intention signal is detected, Lpump is set to zero, representing that the virtual liquid level at the pump station manual end is in an empty state; when a pump station manual pressurization intention signal is detected, Lpump is set to its full-scale value, which is the same as the full-scale value of the virtual fluid dynamic liquid level function L(t), representing that the virtual liquid level at the pump station manual end is instantaneously filled to an overflow state. Virtual pressure potential energy difference. The calculation formula is: =Lpump−L(t), in the formula, Lpump represents the virtual pressure potential energy difference, and L(t) is the pump station operator's pressurized liquid level. The physical meaning of the formula lies in simulating the liquid level difference at both ends of a communicating vessel: when the pump station operator intends to pressurize but has not yet completed preparation, Lpump is at full scale while L(t) is lower; the difference between the two... A positive value indicates that the liquid level at the pump station handpiece is higher than the liquid level at the operator's handpiece. If pressurization is performed at this time, it is equivalent to applying pressure to the high-level end of the communicating vessel. The pressure will be transmitted to the low-level end through the virtual fluid, causing a liquid backflow effect. This physical metaphor corresponds to the risk of the wrench slipping or hand being caught when the pump station operator applies pressure prematurely before the operator is ready in actual operation. When the operator is ready and in a stable physiological state, L(t) is at a high level. If the pump station operator intends to apply pressure at this time, both Lpump and L(t) are at a high level, and the difference between the two is... A value close to zero or even negative indicates that the liquid levels at both ends of the communicating vessel are balanced or the liquid level at the operator's end is higher. In this case, the pressure transmission during the pressurization operation is smooth, which corresponds to the safe pressurization condition when the operator is ready in actual operation.

[0084] The trigger condition for the visual alert is the virtual pressure potential energy difference. Greater than zero. When When the value is greater than zero, the system renders a red inverted triangle warning symbol and a liquid backflow animation at the center of the AR display device interface on the pump operator's training terminal. The red inverted triangle warning symbol uses a high-contrast red fill with a white border. The downward-pointing triangle shape has a psychological suggestion of instability and danger in human visual perception. The warning symbol is positioned in the center of the interface rather than the periphery to ensure that the pump operator can immediately notice the warning when they intend to pressurize. The liquid backflow animation is rendered using a fluid particle system to simulate the visual effect of liquid pouring from a high level to a low level. The flow direction of the particles is from the pump operator's end to the operator's end, and the flow velocity and virtual pressure potential energy difference are used. The value is directly proportional to the value. The larger the pressure, the more severe the liquid backflow, visually demonstrating the physical consequences of improper pressurization leading to pressure backlash. The warning symbol and animation continue to appear until the pump operator removes their finger or their hand condition improves. Until it drops below zero.

[0085] Step S23 adds an active early warning mechanism to the passive state presentation of steps S21 and S22. When the pump operator intends to pressurize but is not yet ready, a visual warning is issued in advance, providing the pump operator with a risk warning of unauthorized operation before actually pressing the button. Capacitive pre-contact detection allows the system to capture the pressurization intention the instant the pump operator's finger touches the button, rather than after the button is pressed, gaining valuable reaction time for the warning. The calculation of the virtual pressure potential energy difference quantitatively compares the operator's state with the pump operator's intention, forming an objective measure of the degree of matching between the two states, avoiding the one-sidedness of judging solely based on one party's state. Without the pre-contact detection and early warning mechanism in step S23, steps S21 and S22 can only present the operator's status information and cannot perceive the pump station operator's intention. The pump station operator may press the button even before the light strip has completely turned green. Although the tactile feedback in the subsequent step S30 can prevent the button from being pressed through a high-resistance state, the tactile feedback occurs at the stage when the finger has already applied force. The early warning in step S23 occurs at the stage when the finger just touches the button. The two form a dual protection of early warning in advance and prevention in the later stage.

[0086] Compared to traditional numerical displays or traffic light signals, the audiovisual fusion presentation method in step S20 offers stronger intuitive perception and cognitive load advantages. Numerical displays require trainees to actively read and interpret the numerical meanings, and while traffic light signals simplify the judgment process, they lose the fine details of intermediate states. The continuous length changes and gradual color textures of the virtual breathing light band retain all the details of state changes, allowing trainees to perceive the light band's state through peripheral vision without conscious interpretation. The continuous playback of background sound effects ensures that state information is still transmitted through the auditory channel even when the trainee's attention is focused on the work object. The auditory entrainment effect induces respiratory synchronization, establishing physiological coordination unconsciously. Pre-contact detection and virtual pressure potential energy difference calculation incorporate the pump operator's operational intentions into the system evaluation scope, enabling the system not only to present the operator's state but also to predict the operator's behavior and issue early warnings, thus upgrading the system from passive display to proactive early warning.

[0087] Step S30: Calculate the virtual fluid viscosity coefficient based on the virtual fluid dynamic liquid surface function and the virtual fluid liquid surface fluctuation rate, convert the virtual fluid viscosity coefficient into a damping control current value for variable damping tactile feedback control, and establish a two-way interactive mechanism for tactile pre-compression tightening and reverse ejection.

[0088] Step S30 extends the collaborative state data packet generated in step S10 and the audiovisual presentation mechanism established in step S20 to the tactile feedback channel, constructing a mapping control system from virtual fluid state to physical damping force. When performing hydraulic bolt tightening operations in the confined space of the wind turbine hub, there is physical isolation between the operator and the pump station operator. Even if the pump station operator obtains the operator's visual state information and auditory rhythm information through step S20, they may still perform pressurization operations before the operator is fully ready due to operational inertia or misjudgment. The visual and auditory channels are feedback mechanisms at the information presentation level and do not have physical constraints on the pump station operator's operational behavior. The pump station operator can ignore the light strip warning and background sound changes and forcibly press the button. Step S30 integrates a magnetorheological fluid damper inside the pump station operator's pressurization controller button, converting the virtual fluid parameters representing the operator's ready state in step S10 into the physical pressing resistance of the button. This allows the pump station operator to directly judge whether it is a suitable time to pressurize by the magnitude of the resistance felt by their fingers when attempting to press the button. When the operator is not yet ready to grip the button or is in a tense physiological state, the button exhibits high resistance, making it difficult for the pump station hand to press. When the operator is ready to grip the button and is in a stable physiological state, the button exhibits low resistance, allowing the pump station hand to press it smoothly. This tactile feedback mechanism establishes a flexible interlocking relationship between the operator's state and the pump station hand's behavior at the physical level. The pump station hand can obtain the operator's readiness state through instinctive force perception without having to consciously interpret visual or auditory information. The two-way interactive mechanism of tactile pre-pressure tightening and reverse ejection further upgrades the state transmission between the operator and the pump station hand from a one-way presentation to a two-way physical connection. When the pump station hand button enters the preparatory stroke, the operator's glove produces a tightening sensation as a pressure warning. When the operator releases the hand during the warning period, the pump station hand button is physically ejected as an emergency brake, forming a real-time tactile interlock that transcends physical isolation.

[0089] Further, step S30 includes:

[0090] Step S31: Normalize the virtual fluid dynamic surface function and the virtual fluid surface turbulence. Calculate the virtual fluid viscosity coefficient based on the normalized virtual fluid dynamic surface function, the normalized virtual fluid surface turbulence, and the preset reference damping constant.

[0091] Specifically, the viscosity coefficient of a virtual fluid is a quantitative parameter characterizing the resistance to fluid flow within a virtual communicating vessel. Its physical metaphor originates from the concept of dynamic viscosity coefficient in real fluid mechanics. In real fluids, the viscosity coefficient determines the frictional resistance between molecules during flow; a higher viscosity coefficient makes fluid flow more difficult, while a lower viscosity coefficient makes fluid flow smoother. Step S31 introduces this physical characteristic by calculating the viscosity coefficient of the virtual fluid based on the operator's readiness and stability. This results in the virtual fluid exhibiting high viscosity when the operator is not ready and low viscosity when the operator is ready. This change in viscosity will be mapped to the physical pressing resistance of the pump station's hand button in step S32.

[0092] The normalization of the virtual fluid dynamic liquid surface function adopts the full-scale normalization method, which divides the virtual fluid dynamic liquid surface function L(t) calculated in step S12 by its full-scale value L. max The normalized virtual fluid dynamic liquid surface function L(t) is obtained. norm Full-scale value L max The determination is based on the mathematical expression of the virtual fluid dynamic liquid level function, when the virtual fluid injection effectiveness index E inject When the full-scale value of 1 is reached and respiration is at the peak of inspiration, causing the sinusoidal modulation term to reach its positive maximum value, the virtual fluid dynamic liquid level function reaches its theoretical maximum value, i.e., L. max =1×(1+A) The normalization of the virtual fluid surface volatility is achieved using a threshold normalization method, which normalizes the virtual fluid surface volatility calculated in step S12. The absolute value divided by the turbulence threshold Normalized virtual fluid surface variability Turbulence threshold The normalization benchmark has a clear physical meaning as follows: when A value less than 1 indicates that the virtual fluid is in a laminar flow state. A value greater than or equal to 1 indicates that the virtual fluid is in a turbulent state. A turbulence threshold, rather than the theoretical maximum value of volatility, is used as the normalization benchmark. The value can directly reflect the relative position of the fluid state from the turbulent boundary. The closer to 0, the smoother the fluid flow. The closer a value is to or greater than 1, the closer the fluid is to or the more turbulent it is.

[0093] Reference damping constant K baseThese are preset system parameters that characterize the basic viscosity characteristics of the virtual fluid under standard conditions. Their physical meaning corresponds to the equivalent viscosity coefficient of the virtual fluid when the operator's hand readiness and stability are at a moderate level. The method for determining the reference damping constant combines the working characteristics of the magnetorheological fluid damper with the pressure sensation characteristics of human fingers: First, the intrinsic damping force of the magnetorheological fluid damper at zero current and the saturation damping force at maximum current are determined; the difference between the two is defined as the damping force adjustment range. Second, through human factors engineering experiments, the threshold of pressure resistance difference that can be clearly distinguished by human fingers is determined, dividing the damping force adjustment range into several distinguishable resistance levels. The reference damping constant is set to the value at the midpoint of the damping force adjustment range corresponding to the virtual fluid viscosity coefficient in the operator's moderate readiness state. For example, if the damping force adjustment range is 5N to 50N, and the midpoint corresponds to 27.5N, then the reference damping constant is set to the value that maps the virtual fluid viscosity coefficient in the moderate readiness state to a damping force of 27.5N.

[0094] Virtual fluid viscosity coefficient η virtual The calculation formula is: , where K base As the numerator, K is the fundamental order of magnitude that determines the viscosity coefficient of the virtual fluid. base The larger the value, the higher the overall viscosity coefficient; L(t) norm The first factor in the denominator reflects the operator's readiness to grip, L(t). norm The larger the value, the more fully the operator holds the virtual liquid and the higher the virtual liquid level, which increases the denominator and thus reduces the viscosity coefficient of the virtual fluid. This results in the virtual fluid becoming thinner and the flow resistance decreasing. The second factor in the denominator reflects the physiological stability of the operator. This factor is maximized using the function max(·) to ensure that its value is always non-negative: when When the value is less than 1, the output of the max function is equal to ( ), A smaller value indicates smoother virtual fluid fluctuations and a more stable operator state, bringing the factor closer to 1. This maintains a larger denominator, thus keeping the virtual fluid viscosity coefficient relatively small, corresponding to a thinner virtual fluid and smoother flow. When the value approaches 1, the output of the max function approaches 0, and the denominator approaches ε, causing the viscosity coefficient of the virtual fluid to increase sharply, corresponding to the virtual fluid tending to solidify and its flow being hindered; when When the value is greater than or equal to 1, it indicates that the virtual fluid has entered a turbulent state, and the max function will ( The negative value of ) is truncated to 0, leaving only the ε term in the denominator, and the virtual fluid viscosity coefficient reaches its maximum value K. base / ε corresponds to the virtual fluid being completely solidified and its flow completely blocked. This truncation ensures that the viscosity coefficient of the virtual fluid is positive under any operating condition, consistent with the physical meaning of the viscosity coefficient. ε is a small zero-prevention term, whose value is set to a positive number much smaller than the normal range of other terms in the denominator. For example, it can be set to 0.001. The introduction of the zero-prevention term ensures that when the virtual fluid is completely solidified and its flow is completely blocked, the viscosity coefficient of the virtual fluid is positive under any operating condition, consistent with the physical meaning of the viscosity coefficient. When the denominator approaches 0, no division-to-zero error will occur, allowing the system to still output an effective virtual fluid viscosity coefficient value under extreme operating conditions.

[0095] The formula's design logic is based on the dual influence of the operator's readiness and stability on the virtual fluid's flow characteristics. Sufficient operator grip keeps the virtual fluid's dynamic surface function high, analogous to a fully filled container, indicating favorable potential energy conditions and low flow resistance. Operator physiological stability keeps the virtual fluid's surface volatility low, analogous to laminar flow, indicating smooth dynamic conditions and low flow resistance. The two factors are multiplied in the denominator, resulting in a multiplicative effect on the virtual fluid viscosity coefficient's response to these conditions: only when the operator's grip is sufficient and stable are both factors at their maximum values, leading to a larger denominator and a smaller virtual fluid viscosity coefficient; if either condition is not met, the corresponding factor decreases, decreasing the denominator and increasing the virtual fluid viscosity coefficient. This multiplicative effect ensures a sensitive response of the virtual fluid viscosity coefficient to the operator's overall readiness, avoiding the risk of misjudgment based on a single indicator. When L(t) norm Approaching 1 and As the value approaches 0, the denominator approaches 1, and the virtual fluid viscosity coefficient approaches K. base This corresponds to the lowest viscous state when the operator is fully ready; when L(t) norm As the value approaches 0, the denominator approaches ε, and the virtual fluid viscosity coefficient approaches K. base The quotient obtained by dividing by ε; when When greater than or equal to 1, The output is 0, the denominator also approaches ε, and the virtual fluid viscosity coefficient approaches K. base The quotient obtained by dividing by ε is much larger than K. base This corresponds to the highest viscosity state when the operator is not fully ready or has entered a turbulent state.

[0096] Step S31 fuses the multidimensional data characterizing the operator's state from step S10 into a single virtual fluid viscosity coefficient, achieving a dimensionality reduction mapping from a two-dimensional state space of readiness and stability to a one-dimensional viscosity characteristic. Normalization eliminates the dimensional and magnitude differences between the virtual fluid dynamic surface function and the virtual fluid surface fluctuation rate, allowing both to participate in viscosity coefficient calculation within a unified numerical range. This avoids the weight imbalance problem where an excessively large value of one parameter masks the influence of another. The multiplicative structure of the denominator makes the virtual fluid viscosity coefficient exhibit nonlinear response characteristics to both normalized parameters. The characteristic of a sharp increase in viscosity coefficient when either parameter deteriorates enhances the system's sensitivity to unsafe operating conditions, ensuring that any insufficient preparation or instability of the operator can be promptly reflected as high viscosity characteristics of the virtual fluid. Without the normalization and viscosity coefficient calculation mechanism in step S31, the magnetorheological fluid damper in step S32 will lack a control signal source. The damper can only maintain a fixed damping state or use simple binary switching, and cannot achieve fine tactile feedback of the damping force changing continuously with the operator's state. The pump station operator will not be able to perceive the transition process from the never-ready to the fully-ready state of the operator through the gradual change of pressing resistance.

[0097] Step S32: Convert the virtual fluid viscosity coefficient into a damping control current value, and use the damping control current value to perform variable damping tactile feedback control.

[0098] Specifically, the pump station manual pressurization controller is an operating interface component in the pump station operator training terminal used to simulate the pressurization switch of an actual hydraulic pump station. Its core component is a pressable physical button. The button of the pump station manual pressurization controller integrates a miniature magnetorheological fluid damper. A magnetorheological fluid damper is an electromechanical device that utilizes the controllable viscosity change characteristic of magnetorheological fluid under the action of an external magnetic field to achieve damping force adjustment. Magnetorheological fluid is a smart fluid material composed of micron-sized ferromagnetic particles suspended in a carrier liquid. In the absence of a magnetic field, the magnetorheological fluid exhibits low-viscosity liquid characteristics, with the ferromagnetic particles freely dispersed in the carrier liquid. When an external magnetic field is applied, the ferromagnetic particles align along the magnetic field lines to form a chain-like structure. The formation of this chain-like structure causes a sharp increase in the yield stress of the magnetorheological fluid, macroscopically manifested as a significant increase in viscosity, transforming the fluid from a readily flowable, thin state to a difficult-to-flow, solidified state. There is a controllable correspondence between the magnetic field strength and the viscosity of the magnetorheological fluid; by adjusting the magnetic field strength applied to the magnetorheological fluid, its viscosity can be continuously adjusted. The structure of the miniature magnetorheological fluid damper consists of three parts: a damping cavity, a piston, and an excitation coil. The damping cavity is a sealed cylindrical cavity filled with magnetorheological fluid. The piston is located at the center of the damping cavity and is connected to the mechanical stroke mechanism of the button. When the button is pressed, the piston moves downward and compresses the magnetorheological fluid inside the damping cavity. The excitation coil is wound around the outer wall of the damping cavity and generates a magnetic field inside the damping cavity when energized. When current is passed through the excitation coil, the viscosity of the magnetorheological fluid increases with the increase of current, and the damping force on the piston increases accordingly, thus increasing the pressing resistance of the button. When the current in the excitation coil decreases or is disconnected, the viscosity of the magnetorheological fluid decreases, the damping force on the piston decreases, and the pressing resistance of the button decreases accordingly.

[0099] Damping control current value I coil This is the current applied to the excitation coil of the miniature magnetorheological fluid damper, and its value directly determines the magnitude of the damping force generated by the damper. Virtual fluid viscosity coefficient η virtual To the damping control current value I coilThe conversion employs a piecewise linear mapping method to adapt to the operating characteristic curve of the magnetorheological fluid damper. The operating characteristic curve of the magnetorheological fluid damper describes the relationship between excitation current and damping force. Influenced by the material properties of the magnetorheological fluid and the design parameters of the excitation coil, this curve typically exhibits nonlinear characteristics. In the low-current range, the damping force increases slowly with increasing current; in the medium-current range, the damping force increases rapidly with increasing current; and in the high-current range, the damping force tends to saturate. The piecewise linear mapping divides the range of the virtual fluid viscosity coefficient into several segments. Within each segment, an independent linear mapping coefficient is used to convert the virtual fluid viscosity coefficient into a damping control current value. The mapping coefficient for each segment is determined based on the inverse function of the damper's operating characteristic curve, resulting in an approximately linear correspondence between the final pressing resistance and the virtual fluid viscosity coefficient. For example, if the virtual fluid viscosity coefficient ranges from 1 to 1000 and the excitation coil operating current ranges from 0 to 2A, the virtual fluid viscosity coefficient can be divided into three segments: 1 to 10, 10 to 100, and 100 to 1000, which correspond to three segments of excitation current: 0 to 0.5A, 0.5 to 1.5A, and 1.5 to 2A, respectively. The damping control current value is calculated using linear interpolation within each segment.

[0100] The tactile feedback control is executed by the real-time control unit of the pump station operator training terminal. The real-time control unit reads the latest virtual fluid viscosity coefficient value from the coordinated state data packet at fixed intervals, calculates the damping control current value through piecewise linear mapping, and outputs the calculated damping control current value to the excitation coil of the miniature magnetorheological fluid damper through a current drive circuit. The effect of the tactile feedback is divided into two typical states: high resistance and low resistance, based on the value of the virtual fluid viscosity coefficient. When the virtual fluid viscosity coefficient η... virtual When the value is large, the damping control current value I coil In the high-value range, the miniature magnetorheological fluid damper generates a large damping force, causing the pump station operator's pressure controller button to exhibit a high-resistance state. Under high resistance, the button's downward pressure resistance increases significantly, requiring the operator to apply considerable finger force to overcome the damper's resistance and displace the button. The tactile experience of this high-resistance state is designed to mimic a hydraulic lock-up sensation, simulating the feel of a pressure-locked valve closing in a real hydraulic system. The operator perceives the button as hydraulically locked and unpressable. This tactile feedback from the high-resistance state psychologically signals inoperability; upon experiencing the hydraulic lock-up sensation and the vibration from particle friction, the operator instinctively stops pressing or reduces the pressure, forming a conditioned reflex to avoid unsafe operating conditions.

[0101] When the virtual fluid viscosity coefficient η virtual When it is small, the damping control current value I coilWhen the resistance is in the low range, even approaching zero, the damping force generated by the miniature magnetorheological fluid damper is minimized, and the pump station manual pressure controller button exhibits a low-resistance state. In this low-resistance state, the button's downward resistance is significantly reduced, requiring only minimal finger pressure from the pump station operator to smoothly press down and trigger the pressure operation. The tactile experience of this low-resistance state is designed as an oily, sucking sensation, simulating the tactile characteristics of a pressure relief valve opening in a real hydraulic system. The pump station operator feels as if the button is smoothly drawn in, as if enveloped in lubricating oil. This oily, sucking sensation is achieved by maintaining the damping control current value at a low level and eliminating any modulation fluctuations. The magnetorheological fluid remains in a low-viscosity, thin state, resulting in almost no resistance or obstruction during piston movement, making the pressing action as smooth as flowing water. The tactile feedback of the low-resistance state psychologically transmits a signal that it can be triggered. After experiencing the oily, sucking sensation, the pump station operator establishes a conditioned association between this sensation and the safe operating timing, forming muscle memory to decisively execute pressure when the low-resistance state is detected.

[0102] The transition between high and low resistance states is not a binary jump but a continuous gradual change. The continuous change in the virtual fluid viscosity coefficient is mapped to a continuous change in pressing resistance through the continuous adjustment of the damping control current value. The pump operator can perceive the gradual trend of resistance from high to low or from low to high during button pressing, and this trend carries real-time information about the operator's state changes. For example, when the operator gradually tightens their grip on the wrench from a relaxed state and adjusts their breathing to a more stable state, the virtual fluid viscosity coefficient gradually decreases, and the pump operator perceives a gradual decrease in button resistance. This trend in resistance change allows the pump operator to anticipate that the operator is about to enter a ready state, thus preparing for pressurization in advance. Conversely, when the operator's grip strength decreases or breathing accelerates due to fatigue or tension, the virtual fluid viscosity coefficient gradually increases, and the pump operator perceives a gradual increase in button resistance. This trend in resistance change allows the pump operator to promptly detect a deterioration in the operator's state, thus proactively delaying the pressurization operation. Compared to binary traffic light signals, the tactile feedback of gradual resistance change has a richer information layer, upgrading the pump station operator's perception of the operator's status from a binary judgment of ready or not ready to a continuous perception of the degree of readyness.

[0103] Step S32 converts the virtual fluid viscosity coefficient calculated in step S31 into a physical resistance change that can be perceived by the pump operator's touch, opening a tactile feedback channel in addition to the audiovisual feedback channel provided in step S20. Compared to the visual and auditory channels, the tactile channel has the characteristic of forced perception. The pump operator can close their eyes and ignore changes in the light strip, or be distracted and ignore changes in background sound effects, but cannot ignore the resistance change felt when pressing the button. Tactile feedback provides an indispensable status cues in work scenarios where attention resources are limited. The use of a magnetorheological fluid damper enables millisecond-level response speed and continuous adjustability of the damping force adjustment. Compared to the binary switching of traditional electromagnetic locking mechanisms, the magnetorheological fluid damper can achieve stepless adjustment of the damping force, establishing a continuous correspondence between the pressing resistance and the operator's state. The pump operator can perceive subtle changes in the operator's state through gradual changes in resistance. The design of a lubricating suction sensation under low resistance enhances the positive excitation effect, enabling the pump operator to quickly and decisively complete the pressurization operation when the operator is ready, avoiding efficiency losses caused by hesitation and delay. Without the variable damping tactile feedback mechanism in step S32, the virtual fluid viscosity coefficient calculated in step S31 will remain only at the numerical level and cannot be perceived by the pump operator. The pump operator's judgment of the operator's state will rely entirely on the visual and auditory channels. Under conditions of distraction or environmental interference, the operator may ignore state changes and perform illegal operations, and the tactile interlocking goal of step S30 will not be achieved.

[0104] Step S33: When the pump station manual pressurization controller button is detected to enter the preparatory stroke, a pre-pressure tightening signal is sent to the operator training terminal, driving the shape memory alloy component of the operator training terminal to generate a tactile tightening sensation; the effective grip total pressure in the collaborative status data packet is monitored in real time, and when a sudden drop in the effective grip total pressure is detected during the effective period of the pre-pressure tightening signal, the reverse ejection mechanism is triggered.

[0105] Specifically, the travel of the manual pressurization controller button in the pump station is divided into two stages: the preparatory travel and the execution travel. The preparatory travel is defined as the range from when the button is pressed down from its initial position until the pressurization command is triggered. The execution travel is defined as the range from when the button is pressed down from the position where the pressurization command is triggered until the end of the travel. The preparatory travel is set based on the safety buffer requirements of the hydraulic system operation. In actual hydraulic pump station operation, after the operator presses the pressurization button, the hydraulic system needs a certain amount of time to build up pressure. The preparatory travel corresponds to the buffer stage where the operator's finger has touched and begun to press the button, but the hydraulic system has not yet responded. The detection of the preparatory travel is achieved through the stroke displacement sensor built into the manual pressurization controller button. The stroke displacement sensor adopts a Hall effect sensor or a grating displacement sensor, which can detect the amount of downward displacement of the button relative to its initial position in real time. For example, if the total travel of the button is set to 10mm, the preparatory travel can be set to the range of 0 to 5mm, and the execution travel to the range of 5mm to 10mm. When the stroke displacement sensor detects that the button's downward displacement enters the 0 to 5mm range, the system determines that the button has entered the preparatory travel state. The pre-pressurization tightening signal is a control signal generated by the pump station operator training terminal and sent to the operator training terminal when the pump station operator pressurization controller button enters the preparatory stroke. The transmission of the pre-pressurization tightening signal relies on the industrial Ethernet architecture established in step S13. The physical significance of the pre-pressurization tightening signal is to inform the operator that the pump station operator has begun the preparatory action of pressurization, providing the operator with a psychological preparation time window for the impending pressurization.

[0106] The training gloves worn by operators in the operator training terminal integrate a shape memory alloy component on the back. Shape memory alloy is a smart metallic material with a shape memory effect, capable of reversible shape changes under temperature variations or electric heating. The shape memory alloy component uses a shrinkable band structure woven from nickel-titanium alloy wires, which wraps around the back of the hand area of ​​the glove. When the shape memory alloy component is not energized, it is soft and flexible, and the shrinkable band is loose, not putting pressure on the back of the hand. When an electric current is applied, the Joule heating generated by the current causes the shape memory alloy component to heat up, resulting in the shrinkage of the alloy wires. The shrinkable band then tightens, creating a circumferential pressure on the back of the hand, providing the operator with a tactile stimulation of the tightened hand.

[0107] After receiving the pre-tightening signal, the operator training terminal drives the control circuit of the shape memory alloy component to supply a heating current to the alloy wire. The alloy wire contracts, tightening the shrink band, and the operator feels a continuous tactile tightening sensation on the back of the glove. The duration of the tactile tightening sensation is determined by the duration of the pre-tightening signal, which remains active while the button is in its pre-travel position, and the corresponding tactile tightening sensation persists. The intensity of the tactile tightening sensation is determined by both the magnitude of the heating current and the pre-tightening force of the shrink band. The design must ensure that the tightening sensation is sufficiently noticeable to attract the operator's attention without causing pain or affecting grip. The physical significance of the tactile tightening sensation lies in simulating the slight vibration that precedes the bolt's tightening. In actual hydraulic bolt tightening operations, the hydraulic wrench vibrates slightly when it begins to build up pressure, and the operator can perceive this incipient vibration through their hand holding the wrench. The tactile tightening sensation transfers this premonitory signal from the wrench end to the glove end, allowing the operator to sense that the pump station has begun pressurizing before the hydraulic pressure is actually built up, thus giving them a final psychological preparation time or an opportunity to withdraw the machine.

[0108] Real-time monitoring of the effective grip total pressure is continuously performed by the pressure sensor array in step S11, and the effective grip total pressure P is collected. total Data is synchronously distributed to the pump station operator training terminal and the operator training terminal along with the coordinated status data packet. The detection of a sudden drop in effective grip total pressure employs a method combining first-order difference and threshold comparison: calculating the change in effective grip total pressure between two adjacent sampling periods. , equal to P in the current period total Subtract P from the previous period total ;Will Compared with the preset drop threshold When comparing, A sudden drop in total effective grip pressure is determined when the value is negative and its absolute value is greater than the drop threshold. The drop threshold is determined based on the statistical characteristics of grip force changes during normal operation. During the system calibration phase, grip force change data of multiple trainees during normal grip and release actions are collected. The amplitude difference between normal grip force fluctuations and grip force drops caused by active release actions is analyzed, and the critical value that can effectively distinguish between the two situations is set as the drop threshold. For example, if the amplitude of grip force fluctuations during normal grip does not exceed 5% of the full scale, while the grip force drop caused by release actions usually exceeds 30% of the full scale, then the drop threshold can be set to 15% of the full scale, so that normal fluctuations do not trigger drop detection, while release actions can reliably trigger drop detection.

[0109] The reverse ejection mechanism is an emergency braking action triggered by the pump operator training terminal when a sudden drop in the effective grip pressure is detected. The execution of the reverse ejection mechanism involves two parallel actions: a damping locking action instantly increases the damping control current value in step S32 to its maximum value, causing the micro magnetorheological fluid damper to generate maximum damping force, locking the pump operator's pressurization controller button in its current position and preventing the button from being pressed further into the execution stroke; the pulse ejection action drives the voice coil motor located below the pump operator's pressurization controller button to generate an upward instantaneous pulse force, physically ejecting the pump operator's fingers. The voice coil motor is a linear reciprocating actuator composed of a permanent magnet and a coil. When current is applied, the coil experiences a Lorentz force in the permanent magnet's magnetic field, resulting in linear motion. The voice coil motor is mounted at the bottom of the button's travel mechanism, and its direction of motion is opposite to the button's downward direction. When a pulse current is applied to the voice coil motor, it generates an upward impact force acting on the bottom of the button, quickly ejecting the button and causing the pump operator's fingers to move upward, thus detaching the fingers from the button surface. The force setting of the pulse ejection must be sufficient to overcome the inertial force of the pump station operator's fingers pressing down and the button's own weight, so that the button can be reliably ejected. At the same time, excessive impact force must be avoided to prevent injury to the pump station operator's fingers.

[0110] The reverse ejection mechanism is triggered only when the total effective grip pressure drops sharply during the effective pre-tightening signal period. Specifically, reverse ejection is triggered only when the button is in the pre-stroke state and the operator simultaneously releases their grip. The logic behind this condition is as follows: an effective pre-tightening signal indicates that the pump operator has begun the preparatory stage of the pressurization action, and the hydraulic system is about to respond and begin pressurization. At this point, the operator releasing their grip indicates that they have abandoned their grip for some reason, such as panic or proactively withdrawing their grip upon discovering a potential hazard. If the pump operator continues to press the button into the execution stroke, it will trigger hydraulic system pressurization. Since the operator has already released their grip, the wrench has lost its grip constraint, and during hydraulic pressurization, the wrench may slip or move abnormally, causing injury. Upon detecting this dangerous combination of conditions, the reverse ejection mechanism immediately prevents the pump operator from continuing the pressurization action. It physically ejects the operator's fingers, preventing them from pressing the button further, thus interrupting the hydraulic pressurization process the instant the operator releases their grip, avoiding the dangerous consequences of hydraulic pressurization without a grip.

[0111] The reverse ejection mechanism establishes a direct physical connection between the operator and the pump station operator. The operator's release action is captured by the system through a sudden drop in total effective grip pressure detection. This data is transmitted to the pump station operator's training terminal via a collaborative status data packet, triggering the reverse ejection mechanism to physically eject the pump station operator's fingers. This direct physical association between the operator's release and the ejection of the pump station operator's button develops into a conditioned reflex-like muscle memory and psychological expectation after repeated training. During training, the pump station operator repeatedly experiences the physical feedback of the button being ejected when the operator's condition is abnormal, gradually developing a subconscious awareness of the operator's condition and a cautious judgment of the timing of pressure application. The operator also repeatedly experiences the causal relationship that the release action immediately prevents the pump station operator from continuing to apply pressure, gradually developing a high regard for grip stability and a clear understanding of collaborative responsibility. This muscle memory and psychological expectation built through physical experience has a more profound behavioral shaping effect than safety regulations conveyed through theoretical instruction, enabling trainees to make correct decisions based on instinct when faced with similar situations in real-world operations.

[0112] Step S33 deeply integrates the effective grip total pressure data collected in step S10 with the tactile feedback channel established in step S32, constructing a two-way physical interaction closed loop between the operator and the pump station operator. The pre-pressure tightening signal allows the operator to receive a tactile warning the moment the pump station operator begins the pressurizing action. This warning transmits the pump station operator's behavioral intention to the operator in real time, upgrading the information flow between the two from one-way state sharing to two-way intention perception. The reverse ejection mechanism allows the operator's release action to immediately interrupt the pump station operator's pressurizing process. This mechanism feeds back the operator's abnormal state to the pump station operator's operation execution level in real time, upgrading the state association between the two from the information level to the action level. The two-way interaction of pre-pressure tightening and reverse ejection forms a tactile interlock that transcends physical isolation. Although the operator and the pump station operator cannot directly see or clearly hear each other, they can perceive each other's behavior and state in real time through the tactile channel. This tactile interlock establishes a physiological tacit understanding and trust foundation between the two. Without the two-way interaction mechanism in step S33, the tactile feedback in step S32 can only achieve one-way transmission from the operator's state to the pump station operator's perception. The pump station operator cannot transmit their own pressurization intention to the operator, and the operator's abnormal state cannot directly prevent the pump station operator from performing the operation. The collaboration between the two remains at the information sharing level rather than the behavior interlocking level, and the state silo problem caused by physical isolation cannot be fundamentally resolved.

[0113] Step S30, while establishing a tactile feedback channel, achieves a functional upgrade from passive state presentation to active behavioral constraint. The audiovisual feedback in step S20 is an information presentation mechanism and does not have physical constraint on the pump operator's actions. The tactile feedback in step S30, however, directly affects the pump operator's experience through changes in damping force. A high-resistance state makes unauthorized operations difficult, while a low-resistance state makes compliant operations smooth, forming a flexible guidance for operational behavior. The reverse ejection mechanism further upgrades from flexible guidance to rigid braking. When a dangerous combination of operating conditions is detected, it directly prevents the pump operator from performing any operation. This rigid braking mechanism ensures that even if the pump operator subjectively ignores the status warning, they can still be objectively prevented from performing dangerous operations. The combination of flexible guidance and rigid braking allows step S30 to both respect the pump operator's operational autonomy and safeguard the safety baseline of collaborative operations, achieving a balance between improving operational efficiency and ensuring safety. The pump station operator senses the other operator's status by changing the resistance of the button and detects any abnormalities by releasing the button. The other operator senses the pump station operator's movements by tightening their glove and ensures proper hydraulic pressure build-up by maintaining a grip. This creates a seamless, unspoken coordination between the two operators. Developing this tactile coordination requires repeated training to form muscle memory. The physical interlocking mechanism provided in step S30 creates the necessary conditions for developing this muscle memory, allowing trainees to repeatedly experience the physical differences between correct and incorrect coordination in a safe training environment. This gradually builds the ability to make correct decisions based on instinct in real-world operations.

[0114] Example 2

[0115] This embodiment, based on Embodiment 1, provides a wind turbine bolt hydraulic tightening training system based on the GWO standard, such as... Figure 5 As shown, it includes:

[0116] State mapping module: used to collect the fingertip pressure distribution matrix, head posture data and respiratory rate of the operator; calculate the virtual fluid injection effectiveness index based on the fingertip pressure distribution matrix and head posture data; generate the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate based on the operator's respiratory rate and the virtual fluid injection effectiveness index;

[0117] Audiovisual feedback module: used to render virtual breathing light strips based on the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate, and generate background sound for fluid flow operation for auditory enhancement display; detect the pump station operator's pressurization intention signal, and generate visual warnings in combination with the virtual fluid dynamic liquid level function;

[0118] Tactile feedback module: It is used to calculate the viscosity coefficient of virtual fluid based on the virtual fluid dynamic liquid surface function and the virtual fluid liquid surface fluctuation rate, convert the virtual fluid viscosity coefficient into the damping control current value for variable damping tactile feedback control, and establish a two-way interactive mechanism for tactile pre-compression tightening and reverse ejection.

[0119] Furthermore, in the state mapping module, the method for calculating the virtual fluid injection effectiveness index includes:

[0120] The effective total grip pressure is obtained by performing an integral calculation on the fingertip pressure distribution matrix;

[0121] Determine whether the head posture data meets the posture conditions and whether the total effective grip pressure meets the grip conditions. When both the posture conditions and grip conditions are met, the total effective grip pressure is normalized based on the set minimum grip force threshold to obtain the virtual fluid injection effectiveness index.

[0122] When either the posture condition or the grip condition is not met, the virtual fluid injection effectiveness index is directly set to zero.

[0123] The virtual fluid dynamic surface function uses the virtual fluid injection effectiveness index as the carrier amplitude and the breathing frequency as the modulation frequency; the virtual fluid surface volatility is obtained by performing second derivative analysis on the virtual fluid dynamic surface function.

[0124] Furthermore, in the audiovisual feedback module, the method for performing virtual breathing light strip rendering includes:

[0125] Based on the length of the virtual breathing light band mapped by the virtual fluid dynamic liquid surface function, the virtual fluid liquid surface volatility is compared with a preset turbulence threshold. When the virtual fluid liquid surface volatility... The absolute value is greater than or equal to the turbulence threshold. When the virtual fluid is determined to be in a turbulent state, the surface undulation rate of the virtual fluid is... The absolute value is less than the turbulence threshold. When the virtual fluid is in a laminar flow state, the color and texture of the virtual breathing light band are determined based on whether the flow is turbulent or laminar.

[0126] The method for generating background noise during fluid flow operations includes:

[0127] The synthesized volume is controlled by the virtual fluid dynamic liquid surface function, and the synthesized timbre is controlled by the virtual fluid liquid surface fluctuation rate, thereby generating background sound for fluid flow operations.

[0128] The methods and systems of this application may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the method is for illustrative purposes only, and the steps of the method of this application are not limited to the order specifically described above, unless otherwise specifically stated.

[0129] In addition, the parts of the technical solutions provided in the embodiments of this application that are consistent with the implementation principles of the corresponding technical solutions in the prior art have not been described in detail, so as to avoid excessive elaboration.

[0130] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A training method for hydraulic tightening of wind turbine bolts based on the GWO standard, characterized in that, The method includes: Collect the fingertip pressure distribution matrix, head posture data, and respiratory rate of the operator; calculate the virtual fluid injection effectiveness index based on the fingertip pressure distribution matrix and head posture data; and generate the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate based on the operator's respiratory rate and the virtual fluid injection effectiveness index. Virtual breathing light strip rendering is performed based on the virtual fluid dynamic liquid level function and the virtual fluid liquid level fluctuation rate, and background sound of fluid flow operation is generated for auditory enhancement display; the pump station operator's pressurization intention signal is detected and a visual warning is generated in combination with the virtual fluid dynamic liquid level function; The viscosity coefficient of the virtual fluid is calculated based on the virtual fluid dynamic surface function and the virtual fluid surface fluctuation rate. The virtual fluid viscosity coefficient is converted into a damping control current value for variable damping tactile feedback control, and a two-way interactive mechanism of tactile pre-compression tightening and reverse ejection is established.

2. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 1, characterized in that, The method for calculating the virtual fluid injection effectiveness index includes: The effective total grip pressure is obtained by performing an integral calculation on the fingertip pressure distribution matrix; Determine whether the head posture data meets the posture conditions and whether the total effective grip pressure meets the grip conditions. When both the posture conditions and grip conditions are met, the total effective grip pressure is normalized based on the set minimum grip force threshold to obtain the virtual fluid injection effectiveness index. When either the posture condition or the grip condition is not met, the virtual fluid injection effectiveness index is directly set to zero.

3. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 2, characterized in that, The head attitude data includes head pitch angle and head yaw angle; The method for determining whether the attitude conditions are met is as follows: both the head pitch angle and the head yaw angle fall within the preset safe operating standard attitude space. The method for determining whether the gripping conditions are met is: the total effective gripping pressure is greater than the minimum gripping force threshold.

4. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 3, characterized in that, The virtual fluid dynamic liquid level function uses the virtual fluid injection effectiveness index as the carrier amplitude and the breathing frequency as the modulation frequency.

5. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 4, characterized in that, The virtual fluid surface fluctuation rate is obtained by performing second derivative analysis on the virtual fluid dynamic surface function.

6. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 5, characterized in that, The method for performing virtual breathing light strip rendering includes: Based on the length of the virtual breathing light band mapped by the virtual fluid dynamic liquid surface function, the virtual fluid liquid surface volatility is compared with a preset turbulence threshold. When the virtual fluid liquid surface volatility... The absolute value is greater than or equal to the turbulence threshold. When the virtual fluid is determined to be in a turbulent state, the surface undulation rate of the virtual fluid is... The absolute value is less than the turbulence threshold. When the virtual fluid is in a laminar flow state, the color and texture of the virtual breathing light band are determined based on whether the flow is turbulent or laminar.

7. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 6, characterized in that, The method for generating background noise during fluid flow operations includes: The synthesized volume is controlled by the virtual fluid dynamic liquid surface function, and the synthesized timbre is controlled by the virtual fluid liquid surface fluctuation rate, thereby generating background sound for fluid flow operations.

8. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 7, characterized in that, The method for generating visual alerts by combining virtual fluid dynamic liquid level functions includes: The manual pressurization intention signal of the pump station is used to determine the manual pressurization liquid level of the pump station. The difference between the manual pressurization liquid level of the pump station and the virtual fluid dynamic liquid level function is calculated to obtain the virtual pressure potential energy difference. The trigger condition for the visualization warning is that the virtual pressure potential energy difference is greater than zero.

9. The wind turbine bolt hydraulic tightening training method based on the GWO standard according to claim 8, characterized in that, The method for establishing a two-way interactive mechanism of tactile pre-compression and reverse ejection includes: When the pump station manual pressurization controller button is detected to enter the pre-travel state, a pre-pressure tightening signal is sent to the operator training terminal, which drives the shape memory alloy component of the operator training terminal to generate a tactile tightening sensation; the effective grip total pressure is monitored in real time, and when a sudden drop in the effective grip total pressure is detected during the effective period of the pre-pressure tightening signal, the reverse ejection mechanism is triggered.

10. A wind turbine bolt hydraulic tightening training system based on the GWO standard, used to implement the wind turbine bolt hydraulic tightening training method based on the GWO standard as described in any one of claims 1-9, characterized in that, The system includes: State mapping module: used to collect the fingertip pressure distribution matrix, head posture data and respiratory rate of the operator; calculate the virtual fluid injection effectiveness index based on the fingertip pressure distribution matrix and head posture data; generate the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate based on the operator's respiratory rate and the virtual fluid injection effectiveness index; Audiovisual feedback module: used to render virtual breathing light strips based on the virtual fluid dynamic liquid level function and virtual fluid liquid level fluctuation rate, and generate background sound for fluid flow operation for auditory enhancement display; detect the pump station operator's pressurization intention signal, and generate visual warnings in combination with the virtual fluid dynamic liquid level function; Tactile feedback module: It is used to calculate the viscosity coefficient of virtual fluid based on the virtual fluid dynamic liquid surface function and the virtual fluid liquid surface fluctuation rate, convert the virtual fluid viscosity coefficient into the damping control current value for variable damping tactile feedback control, and establish a two-way interactive mechanism for tactile pre-compression tightening and reverse ejection.

Images