Self-leveling air gap equalization and multi-cavity stable voltage ultra-low disturbance air cavity system and control method

Abstract

The application discloses a self-leveling air gap equalization and multi-cavity stable voltage ultra-low disturbance air cavity system and a control method, and belongs to the field of aircraft control and ground simulation. The system comprises an outer shaft air floating sleeve, an inner shaft movement sleeve, a lifting device, a surrounding air gap monitoring device, a distributed air pressure sensor array and an air path control board card; an automatic air gap pressure difference elimination is realized by adopting a surrounding communication air path, and a partitioned air supply and a feedforward pressure field shaping are matched, and a constant air cavity volume is maintained by the synchronous lifting device to weaken the hysteresis effect. The control method relies on a neural network offline training to build a compensation model, and realizes closed loop stable voltage and disturbance suppression by adopting a feedforward neural backstepping control. The application can effectively improve the precision, stability and dynamic response of vertical microgravity simulation, and solves the multi-source disturbance problem caused by air gap loss, air supply and exhaust coupling and hysteresis feedback of a traditional pneumatic compensation system.

Description

Technical Field

[0001] This invention relates to a self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system and control method, belonging to the field of aircraft control and ground simulation. Background Technology

[0002] With its comprehensive advantages in test duration, simulation accuracy, and engineering applicability, the air-bearing test system has become a core technical means for constructing microgravity and low-friction space environments on the ground. Currently, the system has accumulated mature technical experience in the fields of three-degree-of-freedom rotation and two-degree-of-freedom translational motion simulation, but research on vertical microgravity motion simulation is still in the exploratory stage.

[0003] Compared to vertical microgravity simulation technologies such as mechanical and electrical control, hydraulics, and constant force springs, cylinder-based pneumatic compensation technology is widely used in air flotation test systems due to its advantages of simple structure, strong load-bearing capacity, and high simulation accuracy. However, existing pneumatic compensation structures still have many shortcomings that urgently need improvement. Specifically, traditional pneumatic compensation systems use a single-point air supply and exhaust mode, which makes it difficult to ensure a uniform distribution of the pressure field inside the air flotation guide and the air chamber, thus causing air gap unevenness. This problem not only generates air flotation fluctuations in the vertical direction but also introduces gravitational eccentricity into the test load. Although some test systems adopt independent air supply schemes for the air flotation guide and the air chamber, supply and exhaust coupling problems occur at the air source interface, leading to system chattering. Therefore, it is urgent to design a distributed air supply system to dynamically monitor and control and eliminate problems such as air gap unevenness and supply and exhaust coupling, thereby achieving feedforward pressure field shaping of the air flotation guide and the cylinder.

[0004] Furthermore, the air pressure response rate of pneumatic devices gradually slows down as the air chamber volume increases. This hysteresis feedback effect is further amplified by problems such as low air pressure sensor sampling rate and monolithic layout, leading to a deterioration in the dynamic performance of the test system. Meanwhile, with the increasing height of pneumatic compensation devices, the supply and exhaust coupling problem of traditional pneumatic compensation systems becomes increasingly prominent. Therefore, it is necessary to optimize the traditional air chamber structure based on a distributed air pressure sensor array, focusing on solving the multi-source disturbance problem caused by factors such as air chamber volume changes, hysteresis feedback, and supply and exhaust coupling. Summary of the Invention

[0005] This invention aims to solve the problem of multi-source disturbances caused by factors such as uneven air gap, hysteresis feedback, and supply and exhaust coupling in traditional cylinders. It proposes a self-leveling air gap equalization and multi-cavity pressure stabilization ultra-low disturbance air chamber system and control method.

[0006] The technical solution of this invention:

[0007] A self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air chamber system includes an outer shaft air-bearing sleeve device, an inner shaft moving sleeve device, a lifting device, an air source supply device, and an air pipe. The inner shaft moving sleeve device is coaxially sleeved inside the outer shaft air-bearing sleeve device. The lifting device is located inside the outer shaft air-bearing sleeve device and below the inner shaft moving sleeve device. An air-bearing guide rail is formed between the outer wall of the inner shaft moving sleeve device and the inner wall of the outer shaft air-bearing sleeve device. A sealed air chamber is formed between the lower end of the inner shaft moving sleeve device and the upper end of the lifting device. The air source supply device provides a pressure-stabilized air source to the system through the air pipe. The system also includes a surround-type air gap monitoring device, a distributed air pressure sensor array, multiple proportional valves, an air path control board, a ground monitoring device, a grating ruler, and a load mounting platform.

[0008] The distributed bar pressure sensor array consists of multiple bar pressure sensors;

[0009] The outer shaft air bearing sleeve device is provided with multiple air bearing sleeve device inlet ports, multiple air bearing sleeve device exhaust ports, and four air gap monitoring ports. The multiple air bearing sleeve device inlet ports are embedded in the side wall of the outer shaft air bearing sleeve device and connected to the corresponding proportional valves through air pipes. The multiple air bearing sleeve device exhaust ports are embedded in the side wall of the outer shaft air bearing sleeve device and connected to the corresponding proportional valves through air pipes. The four air gap monitoring ports are respectively located on the upper side of each wall of the outer shaft air bearing sleeve device.

[0010] The upper end of the inner shaft moving sleeve device is smoothly and fixedly connected to the load mounting platform. The inner shaft moving sleeve device is provided with four independent air gap compensation interfaces and multiple throttling holes. The four air gap compensation interfaces are respectively located on the upper side of each wall of the inner shaft moving sleeve device and connected to the corresponding proportional valve through air pipes. The multiple throttling holes are evenly embedded in each wall of the inner shaft moving sleeve device. The air source input through the air gap compensation interface flows out through the throttling holes to the air float guide rail to form a supporting air gap.

[0011] The lifting device includes a moving piston, a lifting mechanism, multiple lifting device air inlets, multiple lifting device exhaust inlets, and a driver. The moving piston is fixedly installed on the upper end of the lifting mechanism. The multiple lifting device air inlets are embedded in the side wall of the moving piston and are connected to the air inlet of the outer shaft air float sleeve device via air pipes. The multiple lifting device exhaust inlets are embedded in the side wall of the moving piston and are connected to the exhaust of the outer shaft air float sleeve device via air pipes. The driver is communicatively connected to the ground monitoring device and is used to drive the lifting mechanism to move the moving piston in a vertical synchronous motion according to the instructions of the ground monitoring device.

[0012] The surrounding air gap monitoring device includes four air gap conduits, a rigid tube, and four air gap monitoring diaphragms. The two ends of the air gap conduits are respectively connected to the air gap compensation interface and the rigid tube. The air gap monitoring diaphragms are embedded inside the rigid tube. The air gap monitoring diaphragms are wirelessly connected to the ground monitoring device to provide real-time feedback on the pressure difference of the air gap at the corresponding position in the air-bearing guide rail.

[0013] The multiple pressure sensors are respectively installed at the air inlet of each air inlet of the outer shaft air inlet device, and are used to collect the air chamber pressure at the corresponding air inlet position in real time. The pressure sensors are connected to the ground monitoring device.

[0014] The multiple proportional valves are respectively connected to the independent closed-loop control loop of the gas path control board. The gas path control board is communicatively connected to the ground monitoring device and is used to convert the digital control commands transmitted by the ground monitoring device into analog control signals to adjust the opening degree and gas flow of the corresponding proportional valves.

[0015] The gas supply device includes multiple pressure-stabilized gas cylinders, which are connected to the inlet of each proportional valve via gas pipes.

[0016] The grating ruler is connected to the ground monitoring device for real-time acquisition of vertical displacement data of the inner shaft moving sleeve device and feedback to the ground monitoring device.

[0017] Specifically, the multiple air inlet ports of the outer shaft air flotation sleeve device are embedded in the right side wall of the outer shaft air flotation sleeve device, and the multiple air outlet ports are embedded in the left side wall of the outer shaft air flotation sleeve device. The number of air inlet ports and air outlet ports are the same and they are set in a one-to-one correspondence.

[0018] Specifically, the lower side of the lifting device is connected to the outside atmosphere, so that the gas pressure on the lower side of the lifting device is always atmospheric pressure.

[0019] Specifically, the ground monitoring device includes a wireless router and an industrial computer. The industrial computer communicates wirelessly with the surround air gap monitoring device through the wireless router and communicates with the air pressure sensor, air path control board, driver, and grating ruler through RS485 serial ports.

[0020] Specifically, the ground monitoring device sends control commands to the driver based on the real-time displacement and speed data of the inner shaft moving sleeve device fed back by the grating ruler, causing the lifting mechanism to drive the moving piston and the inner shaft moving sleeve device to perform vertical movements at the same speed and in the same direction, so as to keep the volume of the air chamber constant; the ground monitoring device adjusts the opening of the proportional valve connected to the corresponding air gap compensation interface based on the air gap pressure difference data fed back by the air gap monitoring diaphragm, until the air gap pressure difference fed back by the air gap monitoring diaphragm is less than a preset threshold, and then stops adjusting; the preset threshold is preset according to different working conditions.

[0021] A self-leveling air gap equalization and multi-cavity voltage-stabilized ultra-low disturbance air cavity control method, based on the aforementioned self-leveling air gap equalization and multi-cavity voltage-stabilized ultra-low disturbance air cavity system, includes the following steps:

[0022] S1. Complete the installation and commissioning of the self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system, and determine the overall mass m of the inner shaft moving sleeve device, load mounting platform and test load;

[0023] S2. Start the self-leveling air gap equalization and multi-cavity pressure stabilizing ultra-low disturbance air cavity system, complete the self-check of system components, and proceed to the next step after confirming that the system is working normally.

[0024] S3. To address the multi-source disturbances of system supply and exhaust coupling, hysteresis feedback, and air film viscous resistance of the air-floating guide rail, an offline training of the neural network is carried out based on the partitioned air supply algorithm. The neural compensation model used for online disturbance compensation is obtained, realizing feedforward pressure field shaping of the air-floating guide rail and air cavity.

[0025] S4. Based on the distributed air pressure sensor array, the partitioned air supply algorithm and the neural compensation model trained in step S3, a vertical microgravity simulation test was conducted. During the test, a feedforward neural backstepping control algorithm was used to perform closed-loop control on the position, speed and air pressure of the inner shaft moving sleeve device. At the same time, the air gap pressure difference in the air-floating guide rail was automatically eliminated through the surrounding connected air path.

[0026] S5. Save the test data and complete the vertical microgravity ground simulation test.

[0027] Specifically, step S3 involves the feedforward pressure field shaping of the air-bearing guide rail and the air cavity, which includes the following steps:

[0028] S3.1 The ground monitoring device sends control signals to initialize the proportional flow of each proportional valve, and uniformly adjusts the input electrical signal of the proportional valve connected to the four air gap compensation interfaces to 10V, and uniformly adjusts the input electrical signal of the proportional valve connected to the air inlet interface and the air outlet interface of the air flotation sleeve device to 0V.

[0029] S3.2 Construct the kinematic and dynamic model of the ultra-low disturbance air cavity system with self-leveling air gap balancing and multi-cavity coordinated pressure stabilization, as shown in the following expression:

[0030] ;

[0031] Where x1 is the vertical displacement of the inner shaft moving sleeve device measured by the grating ruler, x2 is the vertical velocity of the inner shaft moving sleeve device, x3 is the average air chamber pressure, S is the cross-sectional area of ​​the inner shaft moving sleeve device, P0 is the atmospheric pressure, f is the air film viscous resistance generated by the air-bearing guide rail, k is the adiabatic coefficient, R is the ideal gas constant, T is the gas temperature, and u is the control signal. It is the acceleration due to gravity. This represents the initial volume of the sealed air chamber;

[0032] S3.3 Construct a zoned air supply algorithm, defining the output range of the control signal as -12≤u≤12, and establishing the mapping relationship between the control signal u and the input signal of the proportional valve connected to the air inlet interface of each air flotation sleeve device, and between the control signal u and the input signal of the proportional valve connected to the exhaust interface of each air flotation sleeve device:

[0033] ;

[0034] ;

[0035] in, This is the input signal for the proportional valve connected to the air inlet of the air flotation sleeve device. To input signals to the proportional valves connected to the exhaust ports of each air flotation sleeve device, the n proportional valves connected to the air inlet ports of the air flotation sleeve devices are numbered sequentially from bottom to top, and the numbering is defined as follows: Operating parameters of proportional valve When in working condition, ,otherwise The number n and number of proportional valves connected to the exhaust port of the air flotation sleeve device. and working status parameters The definition method is the same as that on the intake side;

[0036] At the same time, the ground monitoring device drives the lifting device and the inner shaft moving sleeve device to move at the same speed according to the real-time data of the grating ruler, so as to keep the air chamber volume constant.

[0037] S3.4. Collect the output data of the control signal u when the inner shaft moving sleeve device is stabilized at different heights, obtain the feedforward dataset for offline training of the neural network, complete the offline training of the neural network based on the feedforward dataset, and construct the neural compensation model.

[0038] Specifically, in step S3.4, the offline training of the neural network includes the following steps:

[0039] S3.4.1 Define the input data vector for neural network training. and output data vector ,in Represents the maximum value of data stored;

[0040] S3.4.2 Adjust the output of the control signal u to control the proportional flow of the corresponding proportional valve, so that the inner shaft moving sleeve device is stabilized in position sequentially. Among them L is the maximum vertical distance traveled.

[0041] S3.4.3, Sequentially record the stable position of the inner shaft moving sleeve device. At that time, the average air chamber pressure collected by the distributed barometric pressure sensor array and the output magnitude of the control signal u ,in For the barometric pressure data collected by the barometric pressure sensor numbered i, the collected data is stored in the input data vector and the output data vector, respectively. and Complete the construction of the feedforward dataset;

[0042] S3.4.4. Conduct offline training of neural networks based on feedforward datasets to construct a neural compensation model for nonlinear mappings.

[0043] Specifically, step S3.4.4 includes the following steps:

[0044] S3.4.4.1. The max-min normalization method is used to process the feedforward dataset. Preprocessing is performed by randomly shuffling the preprocessed dataset and dividing it into training, validation, and test sets in an 8:1:1 ratio;

[0045] S3.4.4.2 Construct a five-layer neural network structure, namely, an input layer, hidden layer 1, hidden layer 2, hidden layer 3, and an output layer; wherein, the input signal of the input layer is a data vector. The data is normalized; the number of neurons in hidden layer 1, hidden layer 2, and hidden layer 3 are set to 16, 32, and 16 respectively, and the ReLU activation function is used. A batch normalization layer is set after each layer, and a Dropout layer is set after hidden layer 2; the output layer uses a linear activation function to output the normalized predicted value of the control signal. ;

[0046] S3.4.4.3, Define the loss function expression as follows:

[0047] ;

[0048] Where B is the batch size; the Adam optimizer is used to achieve adaptive updates of network parameters, the initial learning rate is set to 0.001, the maximum number of training epochs is set to 500, and a gradient pruning strategy is used to suppress gradient explosion. The normalized actual control signal;

[0049] During training, the training set data is divided into several batches according to batch size B in each round, and the batch data is input into the neural network in sequence. The predicted value is calculated and the loss function is solved through forward propagation. Based on the gradient of the loss function, the weights and biases of each layer are updated through the backpropagation algorithm. After each round, the validation set data is input into the network to calculate the loss value and monitor its change. If the loss of the validation set decreases less than the expected set value for 10 consecutive rounds, the training is stopped early to avoid overfitting. If the convergence condition is not met, the network parameters are reinitialized and training is carried out again.

[0050] S3.4.4.4 After training, the test set data is input into the optimal model to complete the final evaluation. The root mean square error, mean absolute error, and coefficient of determination are calculated to verify the generalization ability of the model. If the error index meets the system control accuracy requirements, the model can be put into use; otherwise, the network structure or hyperparameters need to be adjusted and retrained.

[0051] Specifically, the vertical microgravity simulation experiment in step S4 includes the following steps:

[0052] S4.1 Initialize the proportional valves by uniformly adjusting the input electrical signals of the proportional valves connected to the four air gap compensation interfaces to 10V; and uniformly adjusting the input electrical signals of the proportional valves connected to the air inlet interfaces and air outlet interfaces of the multiple air flotation sleeve devices to 0V.

[0053] S4.2 Construct a feedforward neural backstepping control algorithm for a self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system, and define the expressions for intermediate variables:

[0054] ;

[0055] in, The desired movement position of the inner shaft moving sleeve device. and For virtual control law, mean air chamber pressure The formula is calculated based on a distributed barometric pressure sensor array: ;

[0056] Virtual control rate and The expression for the control signal u is:

[0057] ;

[0058] in, , and For the control parameters to be designed, The neural compensation model obtained from step S3;

[0059] S4.3 In each control cycle, the ground monitoring device calculates the control signal u based on the feedforward neural backstepping control algorithm, and processes the solved input signal based on the zoned gas supply algorithm. Output signal The signal is sent to the pneumatic control board to achieve closed-loop control of the position, speed, and air pressure of the inner shaft moving sleeve device; at the same time, the lifting device is driven to move synchronously with the inner shaft moving sleeve device to maintain a constant air volume.

[0060] S4.4 During the closed-loop motion control process, the pressure difference of each air gap in the air-bearing guide rail is collected in real time by the surrounding air gap monitoring device. The ground monitoring device adjusts the input electrical signal of each proportional valve connected to the air gap duct individually according to the pressure difference feedback data to realize independent closed-loop control of the air gap pressure. When the air gap pressure difference fed back by the air gap monitoring diaphragm is less than the preset threshold, the control signal is stopped. The preset threshold is set in advance according to different working conditions.

[0061] The beneficial effects of this invention are:

[0062] This invention effectively solves the problems of air gap unevenness and supply-exhaust coupling in traditional aerodynamic compensation systems by using a distributed air pressure sensor array, a zoned air supply algorithm, a surrounding interconnected air path, and a feedforward pressure field shaping technology, achieving automatic air gap leveling and uniform pressure field distribution. It maintains a constant air chamber volume by relying on a synchronously moving lifting device, significantly reducing the hysteresis feedback effect caused by changes in air chamber volume and significantly improving the system's dynamic response speed. Combined with offline neural network training and feedforward neural backstepping control, it achieves precise compensation and closed-loop suppression of multi-source disturbances, ultimately enabling the air flotation test system to possess ultra-low disturbance, high stability, and high precision characteristics in vertical microgravity simulations, while also enhancing the system's robustness and engineering applicability. Attached Figure Description

[0063] Figure 1 This is a schematic diagram of the system structure of the present invention;

[0064] Figure 2 A schematic diagram of the structure of the outer shaft air-bearing sleeve device and the surrounding air gap monitoring device;

[0065] Figure 3 A schematic diagram of the inner shaft moving sleeve device;

[0066] Figure 4 This is a schematic diagram of the lifting device structure;

[0067] Figure 5 This is a flowchart of a feedforward neural backstepping control method. Detailed Implementation

[0068] Example 1:

[0069] This embodiment provides a self-leveling air gap equalization and multi-cavity pressure stabilization ultra-low disturbance air cavity system, including an outer shaft air float sleeve device 1, an inner shaft motion sleeve device 2, a lifting device 3, a surrounding air gap monitoring device 4, multiple air pressure sensors 5, multiple proportional valves 6, an air circuit control board 7, a ground monitoring device 8, an air source supply device 9, a grating ruler 10, a load mounting platform 11, and an air pipe 12.

[0070] The outer shaft air bearing sleeve device 1 includes multiple outer shaft air bearing sleeve device inlet ports 1-1, multiple outer shaft air bearing sleeve device exhaust ports 1-2, and four air gap monitoring ports 1-3. The multiple outer shaft air bearing sleeve device inlet ports 1-1 are embedded in the right wall of the outer shaft air bearing sleeve device 1 and are connected to the air pressure sensor 5 and the proportional valve 6 through air pipes 12. The multiple outer shaft air bearing sleeve device exhaust ports 1-2 are embedded in the left wall of the outer shaft air bearing sleeve device 1 and are connected to the proportional valve 6 through air pipes 12. The four air gap monitoring ports 1-3 are embedded on the upper side of each wall of the air bearing sleeve device 1.

[0071] The inner shaft moving sleeve device 2 includes four air gap compensation interfaces 2-1 and multiple throttling orifices 2-2. The four air gap compensation interfaces 2-1 are embedded in the upper side of each wall of the inner shaft moving sleeve device 2 and are connected to the proportional valve 6 through air pipes 12. The multiple throttling orifices 2-2 are uniformly embedded in each wall of the inner shaft moving sleeve device 2.

[0072] Furthermore, the space formed by the inner sidewall of the outer shaft air-bearing sleeve device 1 and the outer sidewall of the inner shaft moving sleeve device 2 is called the air-bearing guide rail. The air source entering the compensation interface 2-1 through the air pipe 12 will flow out from the throttle hole 2-2, thereby forming an air gap in the air-bearing guide rail to reduce frictional resistance. The four air gap compensation interfaces 2-1 are independent of each other, so the pressure of the corresponding air gap can be independently controlled by adjusting the proportional flow rate of the corresponding proportional valve 6.

[0073] Furthermore, the space formed by the lower side of the inner shaft moving sleeve device 2 and the upper side of the lifting device 3 is called the air chamber. The air pressure in the air chamber is mainly controlled by adjusting the proportional valves of the air inlet 1-1 and the exhaust 1-2 of the outer shaft air float sleeve device; the air gap pressure is mainly controlled by adjusting the proportional valve of the compensation interface 2-1.

[0074] Furthermore, the air supply and exhaust coupling phenomenon can be summarized as follows: a small portion of the air source entering the air chamber will flow out from the junction of the air chamber and the air flotation guide rail, and also from the junction of the air chamber and the lifting device 3; a small portion of the air source entering the air gap will flow out from the junction of the air chamber and the air flotation guide rail, and also from the junction of the air flotation guide rail and the upper atmospheric environment.

[0075] The lifting device 3 includes a moving piston 3-1, a lifting mechanism 3-2, multiple lifting device air inlets 3-3, multiple lifting device exhaust inlets 3-4, and a driver 3-5. The moving piston 3-1 is installed above the lifting mechanism 3-2. The lifting mechanism 3-2 and the driver 3-5 are fixedly connected. The multiple lifting device air inlets 3-3 are embedded in the right side of the moving piston 3-1 and connected to the air inlet 1-1 of the outer shaft air-bearing sleeve device through an air pipe 12. The multiple lifting device exhaust inlets 3-4 are embedded in the left side of the moving piston 3-1 and connected to the exhaust inlet 1-2 of the outer shaft air-bearing sleeve device through an air pipe 12. The driver 3-5 communicates with the ground monitoring device 8 via an RS485 serial port and drives the lifting mechanism 3-2 to move vertically according to the command signal from the ground monitoring device 8. The lower side of the lifting device 3 is connected to the outside atmosphere, that is, the gas pressure on the lower side is atmospheric pressure.

[0076] The surrounding air gap monitoring device 4 includes four air gap conduits 4-1, a rigid tube 4-2, and four air gap monitoring diaphragms 4-3. The two ends of the air gap conduits 4-1 are connected to the air gap compensation interface 2-1 and the rigid tube 4-2, respectively. The air gap monitoring diaphragms 4-3 are embedded in the rigid tube 4-2 and communicate with the ground monitoring device 8 via a wireless network to provide real-time feedback on the pressure difference between the two sides of the air gap.

[0077] The multiple air pressure sensors 5 are mainly used to measure the air chamber pressure at the air inlet 1-1 of the outer shaft air bearing sleeve device. They communicate with the ground monitoring device 8 via RS485 serial port to provide real-time feedback of the air chamber pressure at each air inlet.

[0078] The multiple proportional valves 6 are mainly used to adjust the input air pressure of the corresponding air path. Each proportional valve 6 can be independently connected to the corresponding circuit of the air path control board 7. The proportional flow rate is adjusted according to the magnitude of the received analog signal in order to achieve precise control of the air chamber pressure.

[0079] The gas path control board 7 consists of multiple identical and independent closed-loop control loops. Each loop converts the control commands transmitted by the ground monitoring device 8 via RS485 serial port into analog signals, thereby adjusting the proportional flow of the proportional valve 6 connected to the signal.

[0080] The ground monitoring device 8 includes a wireless router and an industrial computer;

[0081] The gas supply device 9 includes multiple gas cylinders and can provide a stable gas source through the gas pipe 12;

[0082] The grating ruler 10 is mainly used to measure the vertical movement distance of the inner shaft moving sleeve device 2. It communicates with the ground monitoring device 8 via RS485 serial port to provide feedback on the real-time position of the moving sleeve device 2.

[0083] The load mounting platform 11 is smoothly connected to the inner shaft motion sleeve device 2, and can carry test loads to simulate microgravity motion.

[0084] The air pipe 12 is mainly used to transmit air source.

[0085] Example 2:

[0086] This embodiment provides a self-leveling air gap equalization and multi-cavity voltage-stabilized ultra-low disturbance air cavity control method, which is based on the system described in Embodiment 1 and includes the following steps:

[0087] S1. Install the self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system, and determine the inner shaft moving sleeve device 2, the load mounting platform 11, and the total mass m of the test load;

[0088] S2. Start the self-leveling air gap equalization and multi-cavity pressure stabilizing ultra-low disturbance air cavity system, and determine whether the outer shaft air float sleeve device 1, inner shaft motion sleeve device 2, lifting device 3, surrounding air gap monitoring device 4, multiple air pressure sensors 5, multiple proportional valves 6, air circuit control board 7, ground monitoring device 8, air source supply device 9, grating ruler 10, load mounting platform 11 and air pipe 12 are working normally. If normal, proceed to the next step.

[0089] S3. To address the multi-source disturbance problems such as supply and exhaust coupling, hysteresis feedback, and air film viscous resistance generated by the air-floating guide rail, an offline training and online compensation of the neural network is performed based on the partitioned air supply algorithm, thereby realizing feedforward pressure field shaping of the air-floating guide rail and the air chamber.

[0090] S4. A vertical microgravity simulation experiment was conducted based on a distributed air pressure sensor array, a zoned air supply algorithm, and a neural compensation model trained in step S3. In the experiment, a feedforward neural backstepping control algorithm was used to perform closed-loop control on the position, speed, and air pressure of the inner axis moving sleeve device 2. At the same time, the air gap pressure difference was automatically eliminated through a surrounding connected air path.

[0091] S5. Save the relevant experimental data and complete the vertical microgravity ground simulation experiment;

[0092] Furthermore, the specific implementation algorithm for the feedforward pressure field shaping of the air-bearing guide rail and air cavity in step S3 includes the following steps:

[0093] S3.1. The ground monitoring device 8 first sends a signal to initialize the proportional flow rate of the proportional valve 6; the range of electrical signal magnitude received by each proportional valve 6 is... The larger the voltage (V), the greater the proportional flow rate of the proportional valve 6 and the greater the input gas pressure. Furthermore, the input voltage of the proportional valve 6 connected to the four air gap compensation interfaces 2-1 is uniformly adjusted to 10V. The input voltage of the proportional valve 6 connected to the air inlet interface 1-1 of the multiple external shaft air float sleeve devices and the exhaust interface 1-2 of the multiple external shaft air float sleeve devices is uniformly adjusted to 0V, that is, there is no supply or exhaust.

[0094] S3.2. Construct the kinematic and dynamic model of the self-leveling air gap equalization and multi-cavity cooperative pressure stabilization ultra-low disturbance air cavity structure system described above:

[0095] ;

[0096] Where x1 is the vertical displacement of the inner shaft moving sleeve device 2 measured by the grating ruler 10, x2 is the vertical velocity of the inner shaft moving sleeve device 2, x3 is the average air chamber pressure, S is the cross-sectional area of ​​the inner shaft moving sleeve device 2, P0 is the atmospheric pressure, f is the air film viscous resistance generated by the air-bearing guide rail, k is the adiabatic coefficient, R is the ideal gas constant, T is the gas temperature, and u is the control signal. It is the acceleration due to gravity. This represents the initial volume of the sealed air chamber.

[0097] S3.3. Construct the aforementioned zoned air supply algorithm, defining the output range of the control signal as -12≤u≤12; further, this control signal and the input signal of each proportional valve 6 connected to the air inlet 1-1 of the outer shaft air float sleeve device. And the input signal of each proportional valve 6 connected to the exhaust port 1-2 of the outer shaft air bearing sleeve device. The relationship between them is:

[0098] ;

[0099] ;

[0100] Where 'n' represents the total number of proportional valves 6 connected to the air inlet 1-1 of the outer shaft air flotation sleeve device, and the proportional valves 6 are numbered sequentially from bottom to top, i.e., the bottommost one is numbered as follows: The top number is ; This indicates whether proportional valve 6, represented by number i, is in the working state. If it is in the working state, then... ,otherwise Furthermore, with the vertical movement of the inner shaft moving sleeve device 2, if and only if the formed air chamber contains the proportional valve 6 numbered i, ,otherwise Furthermore, the total number of proportional valves 6 connected to the exhaust ports 1-2 of the outer shaft air-bearing sleeve device is also n, with their numbers and parameters... The definition method is the same as above;

[0101] Furthermore, when the inner shaft moving sleeve device 2 moves vertically, the ground monitoring device 8 drives the lifting device 3 to work based on the real-time data fed back by the grating ruler 10, so that the movement speed of the lifting device 3 is consistent with the movement speed of the inner shaft moving sleeve device 2, thereby ensuring that the air chamber volume remains unchanged; furthermore, this method reduces the hysteresis feedback effect generated when the air chamber volume increases, effectively improving the dynamic performance of the system.

[0102] S3.4. Acquire the output magnitude of the control signal u when the inner shaft moving sleeve device 2 is stabilized at different heights, thereby obtaining feedforward data for offline training of the neural network. The specific acquisition steps are as follows:

[0103] S3.4.1. Define the input data vector used for neural network training. and output data vector ,in Represents the maximum value of data stored;

[0104] S3.4.2. Adjust the output of the control signal u to control the proportional flow of the proportional valve 6, so that the inner shaft moving sleeve device 2 can be stabilized in position sequentially. Among them L is the maximum vertical distance traveled.

[0105] S3.4.3. Sequentially record the stable position of the inner shaft moving sleeve device 2. At that time, the average air chamber pressure collected by the air pressure sensor 5 and the output magnitude of the control signal u ,in The air pressure data collected by the air pressure sensor numbered i follows the same numbering method as proportional valve 6; then, the above data is stored in the input and output data vectors, i.e. and ;

[0106] S3.4.4. Offline training is conducted using neural networks to construct a high-precision nonlinear mapping model. The specific training steps are as follows:

[0107] S3.4.4.1. Apply the min-max normalization method to the dataset. The dataset is preprocessed, then randomly shuffled and divided into training, validation and test sets in a preset ratio of 8:1:1, which are used for network parameter iteration, training process monitoring and final generalization evaluation, respectively.

[0108] S3.4.4.2. The neural network structure is divided into five layers: input layer, hidden layer 1, hidden layer 2, hidden layer 3, and output layer; furthermore, the input signal of the input layer is a data vector. The data is normalized; the number of neurons in hidden layer 1, hidden layer 2, and hidden layer 3 are set to 16, 32, and 16 respectively, and the ReLU activation function is used. A batch normalization layer is added after each layer to improve training stability, and a Dropout layer is inserted after hidden layer 2 to suppress model overfitting; the output layer uses a linear activation function to directly output the normalized control signal prediction value. ;

[0109] S3.4.4.3. To measure the predictive control signal and the normalized actual control signal The difference between them is defined by the loss function as:

[0110] ;

[0111] Where B is the batch size; furthermore, the Adam optimizer is selected to achieve adaptive updates of network parameters, the initial learning rate is set to 0.001, the maximum number of training epochs is set to 500, and a gradient pruning strategy is adopted to avoid gradient explosion during training. The normalized actual control signal;

[0112] In each training round, the training set data is divided into several batches according to the batch size B, and the batch data is input into the neural network in sequence. The predicted value is calculated and the loss function is solved through forward propagation. Based on the gradient of the loss function, the weights and biases of each layer are updated through the backpropagation algorithm. After each round, the validation set data is input into the network to calculate the loss value and monitor its change. If the loss of the validation set decreases less than the expected set value for 10 consecutive rounds, the training is stopped early to avoid overfitting. If the convergence condition is not met, the network parameters are reinitialized and the training is repeated.

[0113] S3.4.4.4. After training is complete, the test set data is input into the optimal model to complete the final evaluation. The root mean square error, mean absolute error, and coefficient of determination are calculated to verify the generalization ability of the model. If the error index meets the system control accuracy requirements, the model can be put into use; otherwise, the network structure or hyperparameters need to be adjusted and retrained.

[0114] Furthermore, the specific implementation algorithm for the vertical microgravity simulation experiment in step S4 includes the following steps:

[0115] S4.1. First, initialize the proportional valve 6 and adjust the input electrical signal of the proportional valve 6 connected to the four air gap compensation interfaces 2-1 to 10V; adjust the input electrical signal of the proportional valve 6 connected to the air inlet interface 1-1 of the multiple external shaft air bearing sleeve devices and the exhaust interface 1-2 of the multiple external shaft air bearing sleeve devices to 0V.

[0116] S4.2. Construct a feedforward neural backstepping control algorithm for the self-leveling air gap equalization and multi-cavity collaborative voltage stabilization ultra-low disturbance air cavity structure system, and define intermediate variables:

[0117] ;

[0118] in, The desired movement position of the inner shaft moving sleeve device 2. and For virtual control law, mean air chamber pressure Calculated based on a distributed barometric pressure sensor array, i.e. ;

[0119] Furthermore, virtual control rate and The control signal u can be expressed as:

[0120] ;

[0121] in, , and For the control parameters to be designed, The training model obtained in step S3;

[0122] S4.3. In each control cycle, the ground monitoring device 8 calculates the control signal u according to step S4.2, and then uses the solved input signal based on the zonal gas supply algorithm in step S3.3. Output signal The information is sent to the pneumatic control board 7, thereby enabling closed-loop control of the position, speed, and air pressure of the inner shaft moving sleeve device 2.

[0123] Furthermore, when the inner shaft moving sleeve device 2 moves vertically, the lifting device 3 follows the same motion mode as in step S3.3 to counteract the hysteresis feedback phenomenon caused by the change in air chamber volume, thereby improving the dynamic characteristics of the test system.

[0124] S4.4. During the closed-loop motion control in step S4.3, the four air gap monitoring diaphragms 4-3 will send real-time pressure differences between different air gaps to the ground monitoring device 8 respectively; further, the ground monitoring device 8 will adjust the input electrical signal of each proportional valve 6 connected to the air gap duct 4-1 individually according to the feedback data, thereby adjusting the corresponding air gap pressure; further, when the air gap pressure difference fed back by the air gap monitoring diaphragm 4-3 is less than a preset threshold, the control signal will stop being sent, otherwise the signal will continue to be sent to adjust the proportional flow of the proportional valve 6; the preset threshold is preset according to different working conditions, and in this embodiment the preset threshold is 2 Pa.

Claims

1. A self-leveling air gap equalization and multi-cavity pressure-stabilizing ultra-low disturbance air chamber system, comprising an outer shaft air-bearing sleeve device (1), an inner shaft moving sleeve device (2), a lifting device (3), an air source supply device (9), and an air pipe (12), wherein the inner shaft moving sleeve device (2) is coaxially sleeved inside the outer shaft air-bearing sleeve device (1), the lifting device (3) is disposed inside the outer shaft air-bearing sleeve device (1) and located below the inner shaft moving sleeve device (2), an air-bearing guide rail is formed between the outer wall of the inner shaft moving sleeve device (2) and the inner wall of the outer shaft air-bearing sleeve device (1), a sealed air chamber is formed between the lower end of the inner shaft moving sleeve device (2) and the upper end of the lifting device (3), and the air source supply device (9) provides a pressure-stabilizing air source to the system through the air pipe (12), characterized in that, It also includes a surround air gap monitoring device (4), a distributed air pressure sensor array, multiple proportional valves (6), an air circuit control board (7), a ground monitoring device (8), a grating ruler (10), and a load mounting platform (11). The distributed barometric pressure sensor array consists of multiple barometric pressure sensors (5); The outer shaft air bearing sleeve device (1) is provided with multiple air bearing sleeve device air inlet ports (1-1), multiple air bearing sleeve device exhaust ports (1-2), and four air gap monitoring ports (1-3). The multiple air bearing sleeve device air inlet ports (1-1) are embedded in the side wall of the outer shaft air bearing sleeve device (1) and connected to the corresponding proportional valve (6) through the air pipe (12). The multiple air bearing sleeve device exhaust ports (1-2) are embedded in the side wall of the outer shaft air bearing sleeve device (1) and connected to the corresponding proportional valve (6) through the air pipe (12). The four air gap monitoring ports (1-3) are respectively located on the upper side of each wall of the outer shaft air bearing sleeve device (1). The upper end of the inner shaft moving sleeve device (2) is smoothly fixedly connected to the load mounting platform (11). The inner shaft moving sleeve device (2) is provided with four independent air gap compensation interfaces (2-1) and multiple throttling holes (2-2). The four air gap compensation interfaces (2-1) are respectively located on the upper side of each wall of the inner shaft moving sleeve device (2) and connected to the corresponding proportional valve (6) through the air pipe (12). The multiple throttling holes (2-2) are evenly embedded in each wall of the inner shaft moving sleeve device (2). The air source input through the air gap compensation interface (2-1) flows out through the throttling hole (2-2) into the air float guide rail to form a supporting air gap. The lifting device (3) includes a moving piston (3-1), a lifting mechanism (3-2), multiple lifting device air inlets (3-3), multiple lifting device exhaust ports (3-4), and a driver (3-5). The moving piston (3-1) is fixedly installed on the upper end of the lifting mechanism (3-2). The multiple lifting device air inlets (3-3) are embedded in the side wall of the moving piston (3-1) and connected to the air inlet (1-1) of the air bearing sleeve device (1) of the outer shaft air bearing sleeve device (1) through the air pipe (12). The multiple lifting device exhaust ports (3-4) are embedded in the side wall of the moving piston (3-1) and connected to the air bearing sleeve device exhaust port (1-2) of the outer shaft air bearing sleeve device (1) through the air pipe (12). The driver (3-5) is connected to the ground monitoring device (8) for communication and is used to drive the lifting mechanism (3-2) to drive the moving piston (3-1) to make synchronous vertical movement according to the instructions of the ground monitoring device (8). The surrounding air gap monitoring device (4) includes four air gap conduits (4-1), a rigid tube (4-2), and four air gap monitoring diaphragms (4-3). The two ends of the air gap conduits (4-1) are respectively connected to the air gap compensation interface (2-1) and the rigid tube (4-2). The air gap monitoring diaphragms (4-3) are embedded inside the rigid tube (4-2). The air gap monitoring diaphragms (4-3) are wirelessly connected to the ground monitoring device (8) for real-time feedback of the pressure difference of the air gap at the corresponding position in the air-bearing guide rail. The multiple air pressure sensors (5) are respectively set at the air inlet (1-1) of each air inlet of the outer shaft air inlet sleeve device (1) to collect the air chamber pressure at the corresponding air inlet position in real time. The air pressure sensors (5) are connected to the ground monitoring device (8) for communication. The multiple proportional valves (6) are respectively connected to the independent closed-loop control circuit of the gas path control board (7) one by one. The gas path control board (7) is connected to the ground monitoring device (8) for communication, and is used to convert the digital control commands transmitted by the ground monitoring device (8) into analog control signals to adjust the opening degree and gas flow of the corresponding proportional valve (6). The gas supply device (9) includes multiple pressure-stabilized gas cylinders, which are connected to the inlet of each proportional valve (6) via gas pipes (12); The grating ruler (10) is connected to the ground monitoring device (8) for real-time acquisition of vertical displacement data of the inner shaft moving sleeve device (2) and feedback to the ground monitoring device (8).

2. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system according to claim 1, characterized in that, The multiple air inlet ports (1-1) of the outer shaft air flotation sleeve device (1) are embedded in the right side wall of the outer shaft air flotation sleeve device (1), and the multiple air outlet ports (1-2) are embedded in the left side wall of the outer shaft air flotation sleeve device (1). The number of air inlet ports (1-1) and air outlet ports (1-2) are the same and they are set in a one-to-one correspondence.

3. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system according to claim 1, characterized in that, The lower side of the lifting device (3) is connected to the outside atmosphere, so that the gas pressure on the lower side of the lifting device (3) is always atmospheric pressure.

4. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system according to claim 1, characterized in that, The ground monitoring device (8) includes a wireless router and an industrial computer. The industrial computer communicates wirelessly with the surround air gap monitoring device (4) through the wireless router and communicates with the air pressure sensor (5), air path control board (7), driver (3-5), and grating ruler (10) through the RS485 serial port.

5. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system according to claim 1, characterized in that, The ground monitoring device (8) sends a control command to the driver (3-5) based on the real-time displacement and speed data of the inner shaft moving sleeve device (2) fed back by the grating ruler (10), so that the lifting mechanism (3-2) drives the moving piston (3-1) and the inner shaft moving sleeve device (2) to make vertical movements at the same speed and in the same direction, so that the volume of the air chamber remains constant; the ground monitoring device (8) adjusts the opening of the proportional valve (6) connected to the corresponding air gap compensation interface (2-1) based on the air gap pressure difference data fed back by the air gap monitoring diaphragm (4-3), until the air gap pressure difference fed back by the air gap monitoring diaphragm (4-3) is less than the preset threshold, and then stops adjusting; the preset threshold is preset according to different working conditions.

6. A method for self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity control, characterized in that, The system, based on any one of claims 1 to 5, features a self-leveling air gap equalization and multi-cavity voltage-stabilized ultra-low disturbance air cavity, and includes the following steps: S1. Complete the installation and commissioning of the self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system, and determine the overall mass m of the inner shaft moving sleeve device (2), load mounting platform (11) and test load; S2. Start the self-leveling air gap equalization and multi-cavity pressure stabilizing ultra-low disturbance air cavity system, complete the self-check of system components, and proceed to the next step after confirming that the system is working normally. S3. To address the multi-source disturbances of system supply and exhaust coupling, hysteresis feedback, and air film viscous resistance of the air-floating guide rail, an offline training of the neural network is carried out based on the partitioned air supply algorithm. The neural compensation model used for online disturbance compensation is obtained, realizing feedforward pressure field shaping of the air-floating guide rail and air cavity. S4. Based on the distributed air pressure sensor array, the partitioned air supply algorithm and the neural compensation model trained in step S3, a vertical microgravity simulation test was conducted. During the test, a feedforward neural backstepping control algorithm was used to perform closed-loop control on the position, speed and air pressure of the inner shaft motion sleeve device (2). At the same time, the air gap pressure difference in the air-floating guide rail was automatically eliminated through the surrounding connected air path. S5. Save the test data and complete the vertical microgravity ground simulation test.

7. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity control method according to claim 6, characterized in that, Step S3 involves the feedforward pressure field shaping of the air-bearing guide rail and the air cavity, specifically including the following steps: S3.1 The ground monitoring device (8) sends a control signal to initialize the proportional flow of each proportional valve (6), and adjusts the input electrical signal of the proportional valve (6) connected to the four air gap compensation interfaces (2-1) to 10V, and adjusts the input electrical signal of the proportional valve (6) connected to the air inlet interface (1-1) and the air outlet interface (1-2) of the air flotation sleeve device to 0V. S3.2 Construct the kinematic and dynamic model of the ultra-low disturbance air cavity system with self-leveling air gap balancing and multi-cavity coordinated pressure stabilization, as shown in the following expression: ； Where x1 is the vertical displacement of the inner shaft moving sleeve device (2) measured by the grating ruler (10), x2 is the vertical velocity of the inner shaft moving sleeve device (2), x3 is the average air chamber pressure, S is the cross-sectional area of ​​the inner shaft moving sleeve device (2), P0 is the atmospheric pressure, f is the air film viscous resistance generated by the air-bearing guide rail, k is the adiabatic coefficient, R is the ideal gas constant, T is the gas temperature, and u is the control signal. It is the acceleration due to gravity. This represents the initial volume of the sealed air chamber; S3.3 Construct a zoned gas supply algorithm, define the output range of the control signal as -12≤u≤12, establish the mapping relationship between the control signal u and the input signal of the proportional valve (6) connected to the air inlet interface (1-1) of each air flotation sleeve device, and the input signal of the proportional valve (6) connected to the exhaust interface (1-2) of each air flotation sleeve device: ； ； in, For the input signal of the proportional valve (6) connected to the air inlet (1-1) of the air flotation sleeve device, To input signals to the proportional valves (6) connected to the exhaust ports (1-2) of each air flotation sleeve device, the n proportional valves (6) connected to the air inlet port (1-1) of the air flotation sleeve device are numbered sequentially from bottom to top, and the numbering is defined as follows: Operating parameters of proportional valve (6) When in working condition, ,otherwise The number n and number of proportional valves (6) connected to the exhaust port (1-2) of the air flotation sleeve device. and working status parameters The definition method is the same as that on the intake side; At the same time, the ground monitoring device (8) drives the lifting device (3) and the inner shaft moving sleeve device (2) to move at the same speed according to the real-time data of the grating ruler (10) to keep the air chamber volume constant. S3.

4. Collect the output data of the control signal u when the inner shaft moving sleeve device (2) is stable at different heights, obtain the feedforward dataset for offline training of the neural network, complete the offline training of the neural network based on the feedforward dataset, and construct the neural compensation model.

8. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity control method according to claim 7, characterized in that, In step S3.4, the offline training of the neural network specifically includes the following steps: S3.4.1 Define the input data vector for neural network training. and output data vector ,in Represents the maximum value of data stored; S3.4.

2. Adjust the output of the control signal u to control the proportional flow of the corresponding proportional valve (6), so that the inner shaft moving sleeve device (2) is stabilized in position sequentially. Among them L is the maximum vertical distance traveled. S3.4.3, Sequentially record the inner shaft movement sleeve device (2) stabilizing at position. At that time, the average air chamber pressure collected by the distributed barometric pressure sensor array and the output magnitude of the control signal u ,in For the barometric pressure data collected by the barometric pressure sensor numbered i, the collected data is stored in the input data vector and the output data vector, respectively. and Complete the construction of the feedforward dataset; S3.4.

4. Conduct offline training of neural networks based on feedforward datasets to construct a neural compensation model for nonlinear mappings.

9. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity control method according to claim 7, characterized in that, Step S3.4.4 specifically includes the following steps: S3.4.4.

1. The max-min normalization method is used to process the feedforward dataset. Preprocessing is performed by randomly shuffling the preprocessed dataset and dividing it into training, validation, and test sets in an 8:1:1 ratio; S3.4.4.2 Construct a five-layer neural network structure, namely, an input layer, hidden layer 1, hidden layer 2, hidden layer 3, and an output layer; wherein, the input signal of the input layer is a data vector. The data is normalized; the number of neurons in hidden layer 1, hidden layer 2, and hidden layer 3 are set to 16, 32, and 16 respectively, and the ReLU activation function is used. A batch normalization layer is set after each layer, and a Dropout layer is set after hidden layer 2; the output layer uses a linear activation function to output the normalized predicted value of the control signal. ; S3.4.4.3, Define the loss function expression as follows: ； Where B is the batch size; the Adam optimizer is used to adaptively update the network parameters, the initial learning rate is set to 0.001, the maximum number of training epochs is set to 500, and a gradient pruning strategy is used to suppress gradient explosion. The normalized actual control signal; During training, the training set data is divided into several batches according to batch size B in each round, and the batch data is input into the neural network in sequence. The predicted value is calculated and the loss function is solved through forward propagation. Based on the gradient of the loss function, the weights and biases of each layer are updated through the backpropagation algorithm. After each round, the validation set data is input into the network to calculate the loss value and monitor its change. If the loss of the validation set decreases less than the expected set value for 10 consecutive rounds, the training is stopped early to avoid overfitting. If the convergence condition is not met, the network parameters are reinitialized and training is carried out again. S3.4.4.4 After training, the test set data is input into the optimal model to complete the final evaluation. The root mean square error, mean absolute error, and coefficient of determination are calculated to verify the generalization ability of the model. If the error index meets the system control accuracy requirements, the model can be put into use; otherwise, the network structure or hyperparameters need to be adjusted and retrained.

10. The self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity control method according to claim 9, characterized in that, Step S4, the vertical microgravity simulation experiment, specifically includes the following steps: S4.1 Initialize the proportional valve (6), and adjust the input electrical signal of the proportional valve (6) connected to the four air gap compensation interfaces (2-1) to 10V; adjust the input electrical signal of the proportional valve (6) connected to the air inlet interface (1-1) of multiple air flotation sleeve devices and the exhaust interface (1-2) of multiple air flotation sleeve devices to 0V. S4.2 Construct a feedforward neural backstepping control algorithm for a self-leveling air gap equalization and multi-cavity pressure-stabilized ultra-low disturbance air cavity system, and define the expressions for intermediate variables: ； in, The desired movement position of the inner shaft moving sleeve device (2) is given. and For virtual control law, mean air chamber pressure The formula is calculated based on a distributed barometric pressure sensor array: ; Virtual control rate and The expression for the control signal u is: ； in, , and For the control parameters to be designed, The neural compensation model obtained from step S3; S4.3 In each control cycle, the ground monitoring device (8) calculates the control signal u based on the feedforward neural backstepping control algorithm, and converts the solved input signal into a signal based on the zoned gas supply algorithm. Output signal Send to the air circuit control board (7) to realize closed-loop control of the position, speed and air pressure of the inner shaft moving sleeve device (2); at the same time, drive the lifting device (3) to move synchronously with the inner shaft moving sleeve device (2) to maintain the constant air chamber volume; S4.4 During the closed-loop motion control process, the pressure difference of each air gap of the air-bearing guide rail is collected in real time by the surrounding air gap monitoring device (4). The ground monitoring device (8) adjusts the input electrical signal of each proportional valve (6) connected to the air gap duct (4-1) according to the pressure difference feedback data to realize independent closed-loop control of the air gap pressure. When the air gap pressure difference fed back by the air gap monitoring diaphragm (4-3) is less than the preset threshold, the control signal is stopped. The preset threshold is set in advance according to different working conditions.

Images