Self-adaptive viscous damping wall based on machine learning

Abstract

The invention belongs to the technical field of damping walls, and particularly relates to a self-adaptive viscous damping wall based on machine learning, which comprises a channel steel box internally loaded with viscous liquid, the channel steel box is arranged between two damper connecting wall piers, a damping body is vertically arranged in the channel steel box, a self-adaptive mechanism is arranged in the damping body in a penetrating manner, and the self-adaptive mechanism is arranged in the damping body in a penetrating manner. And tackifying mechanisms are arranged at the two ends of the self-adaptive mechanism. According to the self-adaptive viscous damping wall based on machine learning, sliding friction is converted into rolling friction through a ball structure, motion resistance and abrasion are remarkably reduced, and motion smoothness and self-balancing capacity of a damping body under complex multi-dimensional vibration are ensured; additional bending moment and stress concentration caused by mismatching of the movement direction and the structural rigidity are effectively avoided, meanwhile, the built-in belleville spring buffer module can absorb huge impact energy during extreme overload, mechanical hard limiting is provided in cooperation with the limiting groove, and dual protection is provided for the damping wall.

Description

Technical Field

[0001] This invention relates to the field of damping wall technology, and more particularly to an adaptive viscous damping wall based on machine learning. Background Technology

[0002] Damped walls, also known as damper walls or energy dissipation walls, are efficient passive control devices used in modern high-rise buildings, long-span bridges, and other important structures to resist wind loads and seismic forces. Their core function is to dissipate the energy input to the structure, reducing vibration and deformation, thereby improving structural comfort and safety.

[0003] However, existing damping walls still have the following significant problems in practical applications: Under moderate or large earthquakes, the temperature of the viscous fluid rises due to continuous shear. If the temperature rise is too rapid or heat dissipation is poor, the viscosity of the viscous fluid will drop significantly, leading to a decrease in damping force and a significant reduction in energy dissipation capacity, affecting structural safety. Furthermore, the damping plates of traditional damping walls are mostly fixed installations. Seismic excitation is inherently multidimensional; in addition to the main vibration direction, strong lateral forces can cause unexpected bending or torsional deformation of the damping plates, which may lead to fatigue damage or even failure under long-term action, affecting service life and reliability. Moreover, existing devices are mostly in passive operation mode, and their damping characteristics are fixed after installation. They cannot be dynamically adjusted according to the real-time intensity, spectral characteristics, and actual structural response of the earthquake, making it difficult to maintain the optimal energy dissipation state in earthquake events of different magnitudes and characteristics, thus limiting the potential for improving their vibration reduction efficiency. Therefore, the present invention addresses the shortcomings of the above-mentioned technical problems. Summary of the Invention

[0004] Based on the aforementioned technical problems, this invention proposes an adaptive viscous damping wall based on machine learning.

[0005] This invention proposes an adaptive viscous damping wall based on machine learning, comprising a channel steel box filled with viscous fluid, the channel steel box being disposed between two damper connecting wall piers, a damping body being vertically placed inside the channel steel box, the upper surface of the damping body being fixedly connected to the lower surface of the upper damper connecting wall pier via embedded parts, an adaptive mechanism being disposed through the interior of the damping body, and adhesion-enhancing mechanisms being disposed at both ends of the adaptive mechanism, wherein the channel steel box is made of high-strength steel, possessing good rigidity and corrosion resistance, and the damping body is made of metal, the surface of which can be treated for rust prevention.

[0006] The damping body reciprocates along the length of the wall pier connected to the damper via a viscous fluid. During the movement, the adaptive mechanism maintains self-balancing of the damping body.

[0007] The adhesive-enhancing mechanism is in full contact with the viscous liquid inside the channel steel box and provides adhesive damping during vibration.

[0008] Preferably, the lower surface of the channel steel box is symmetrically and hingedly fitted with lifting hydraulic cylinders. The piston rod end of the lifting hydraulic cylinder is hingedly fitted to the upper surface of the lower damper connecting wall pier through another embedded part. The lower surface of the channel steel box is also fixedly fitted with multiple buffer telescopic rods for supporting the channel steel box through several other sets of embedded parts. The lower surface of the buffer telescopic rods is also fixedly fitted to the upper surface of the lower damper connecting wall pier through corresponding embedded parts.

[0009] Through the above technical solution, in order to adapt to the target damping force, the viscosity depth needs to be adjusted. The lifting hydraulic cylinder and the buffer telescopic rod work together to adjust the height of the channel steel box in real time during an earthquake to adapt to different vibration intensities. At the same time, the buffer telescopic rod provides elastic support to avoid rigid impact and extend the service life of the device.

[0010] Preferably, the damping body is composed of a transverse connecting top plate and a vertical adhesive plate fixedly connected to the lower surface of the connecting top plate. The upper surface of the connecting top plate is fixedly connected to the lower surface of the upper damper connecting wall pier through a pre-embedded part. There is a flow gap between the peripheral surface and the lower surface of the adhesive plate and the inner wall of the channel steel box.

[0011] Through the above technical solution, in order to achieve the damping effect brought about by the flow of viscous liquid, a flow gap of 3-5mm is left between the viscous plate and the inner wall of the channel steel box, so that the viscous liquid can form a stable flow field during vibration, enhance the energy dissipation effect, and connect the top plate and the upper wall pier rigidly, directly transfer the load, and have strong structural integrity.

[0012] Preferably, the adaptive mechanism includes through holes arranged in a rectangular array on the surface of the adhesive plate, an outer sleeve fixedly fitted to the inner surface of the through holes, elliptical grooves arranged in a ring array on the inner surface of the outer sleeve, an inner sphere with an inner hole inside the outer sleeve, and a support shaft passing through both sides of the adhesive plate fixedly connected to the inner surface of the inner hole of the inner sphere.

[0013] Through the above technical solution, to achieve self-adaptation between the damping body and the viscous fluid, the outer sleeve serves as a fixed component, its outer wall being fixed to the through hole in the viscous plate via interference fit or welding. The inner sphere serves as a movable component, its inner hole being fixed to the support shaft via key connection or press fit, making the support shaft and the inner sphere an integral unit. Both ends of the support shaft extend out of the viscous plate for connecting to the external adhesion-enhancing mechanism. Because there is an assembly gap between the inner sphere and the inner wall of the outer sleeve, and the inner sphere is not completely locked within the outer sleeve, this gives the joint two rotational degrees of freedom. When the damping body... When moving in a viscous fluid and subjected to fluid resistance or lateral inertial forces from different directions, the support shaft and inner spherical assembly can undergo minute spherical rotation relative to the fixed outer sleeve and viscous plate assembly. This minute rotation capability allows the thickening mechanism connected to both ends of the support shaft to adaptively fine-tune the angle according to the direction of the flow force, thereby more effectively cutting into the viscous fluid flow, optimizing the damping force generation efficiency, and avoiding additional bending moments and stress concentrations caused by the mismatch between the direction of movement and the structural rigidity, significantly improving the smoothness of movement and the structural durability of the damper.

[0014] Preferably, the adaptive mechanism further includes a ball cage slidably connected to the inner surface of the outer sleeve, the inner surface of the ball cage slidingly contacting the outer surface of the inner sphere, the outer surface of the inner sphere having grooves distributed in a ring array corresponding one-to-one with the elliptical grooves, the outer surface of the ball cage having a plurality of windows evenly distributed circumferentially through it, each window rollingly receiving a ball, the elliptical grooves being arranged opposite to the grooves and squeezing the ball, so that the ball rolls in contact with the inner surfaces of the elliptical grooves and the grooves.

[0015] Through the above technical solution, in order to achieve the force transmission path and motion constraint, the force and motion between the inner ball and the outer tube are no longer transmitted through direct sliding friction, but through the intermediate medium of the ball. The ball is in contact with the groove on the inner ball and the elliptical groove on the outer tube at the same time. The geometry of the elliptical groove generates a pre-constraint force on the ball. When the inner ball attempts to rotate relative to the outer tube, the ball will roll purely under the guidance of the groove and the elliptical groove. This mechanism of converting sliding friction into rolling friction greatly reduces the frictional resistance and wear inside the joint, ensuring the sensitivity and low hysteresis of the adaptive motion. The role of the ball cage is to ensure that all the balls always maintain the correct relative position and spacing, prevent them from colliding with each other or falling off, and make the load evenly distributed among the multiple balls.

[0016] Preferably, a conical dust cover is fixedly connected to the outer cylindrical surface of the outer sleeve opening end by a metal clamp, and the other end of the dust cover is fastened to the outer surface of the support shaft by another metal clamp. The dust cover is made of thermoplastic rubber and forms a sealed flexible cavity between the outer sleeve and the support shaft. Lubricating grease is provided in the flexible cavity.

[0017] Through the above technical solution, in order to achieve dynamic sealing of the adaptive mechanism, the conical or bellows shape of the dust cover provides it with axial and radial expansion and contraction capabilities. When the support shaft drives the inner ball to rotate, the relative position between the support shaft and the outer sleeve will change slightly. The flexible dust cover can be stretched, compressed or bent accordingly to maintain a sealed state without generating excessive constraint force or fatigue cracking. The grease in the flexible cavity provides long-term lubrication for the contact surfaces of the balls, grooves and ball cages, and isolates corrosive components that may penetrate the viscous liquid. At the same time, the grease itself has a certain viscosity, which can play an additional buffering and damping role for the small vibrations of the joint.

[0018] Preferably, the adaptive mechanism further includes a force-bearing ring fixedly sleeved on the outer surface of the middle part of the outer sleeve, and a limiting groove communicating with the through hole is provided inside the adhesive plate. The inner surface of the limiting groove is movably sleeved with the outer surface of the force-bearing ring. The adhesive plate also has symmetrically distributed buffer grooves inside. The outer end of the buffer groove is communicating with the through hole and the limiting groove. A butterfly spring is fixedly connected to one inner wall of the buffer groove. The free end of the butterfly spring wraps around and fixes one outer surface of the force-bearing ring.

[0019] Through the above technical solution, the internal components of the adaptive mechanism, such as the outer sleeve and inner ball, are made of alloy steel or wear-resistant cast iron, which can withstand high temperatures of over 200°C. At the same time, in order to avoid the impact on the adaptive mechanism under extreme overload conditions and reduce its service life, when encountering extremely large vibrations far exceeding the design standards, the axial force acting on the support shaft and inner ball assembly may be abnormally large. At this time, the huge axial force will be transmitted to the force ring through the ball and the outer sleeve. The force ring will absorb a large amount of impact energy through the deformation process of the disc spring. The limiting groove is a cavity on the viscous plate with a diameter larger than that of the through hole, allowing the force ring to have a small axial movement space in it. It is connected to the through hole and the buffer groove, so even if the viscous liquid penetrates into the buffer groove and comes into contact with the disc spring, its deformation process will force the viscous liquid to adhere to the surface of the disc spring, thereby playing an additional buffering and damping role.

[0020] Preferably, the adhesion-enhancing mechanism includes wing plates fixedly connected to the outer surfaces of the two free ends of the support shaft. The two wing plates are symmetrically distributed around the center point of the support shaft at a 180-degree rotation. The outer edges of the wing plates are rounded. The upper and lower surfaces of the wing plates are provided with regularly arranged spherical grooves. Adjacent spherical grooves are fixedly connected by cylindrical grooves, thereby forming a grid-like network of interconnected grooves on the upper and lower surfaces of the wing plates.

[0021] Through the above technical solution, the wingplate of the viscosity-enhancing mechanism is made of aluminum alloy or stainless steel, with a smooth surface and shear resistance. To increase the lateral contact area and interaction strength with the viscous fluid, thereby significantly improving damping performance, the wingplate is designed as a smooth, full-bodied sheet structure. Its large surface area provides basic planar contact. The two wingplates are arranged symmetrically at 180 degrees, ensuring that during the reciprocating motion of the damping body, regardless of the direction of motion, at least one wingplate can interact with the viscous fluid at a better angle of attack, providing symmetrical and balanced damping force. The spherical grooves serve as local depressions, and the cylindrical grooves as connecting channels. When the wingplate moves in the viscous fluid, the viscous fluid flows over its surface. The network of connecting channels breaks the stable development of the laminar boundary layer on the flat plate, forcing the fluid to enter, pass through, and exit the various spherical and cylindrical grooves. This process continuously generates microscale eddies, deflection, and reattachment, greatly increasing the shear deformation rate and velocity gradient within the fluid. The rotation function of the adaptive mechanism allows the airfoil to adaptively adjust its angle relative to the incoming flow based on the instantaneous flow field direction. This dynamic adjustment ensures that the airfoil, based on the reinforcement effect of the network groove, can always enter the flow field with a better hydrodynamic attitude, maximizing its efficiency in driving the viscous fluid and generating shear force, and avoiding performance loss or vibration coupling problems that may be caused by a fixed angle. The airfoil is preferably made of high-strength aluminum alloy or austenitic stainless steel, meeting the working temperature range of -30℃ to 150℃, to adapt to the temperature rise that the viscous fluid may generate due to repeated shearing during the earthquake, as well as the temperature changes of the external environment where the building is located, ensuring structural stability and fatigue life.

[0022] Preferably, the viscous damping wall includes a sensor system and an intelligent control system, wherein the sensor system includes:

[0023] Displacement and acceleration sensors mounted on the damping body are used to monitor the displacement, velocity, and acceleration of the damping body in the viscous fluid in real time.

[0024] A temperature sensor is installed on the inner wall of the channel steel box.

[0025] A pressure sensor is mounted on the surface of the viscous plate.

[0026] Strain sensor installed on the damper connecting wall pier.

[0027] The intelligent control system includes:

[0028] The data acquisition module connects to all sensors and is used to collect sensor data in real time.

[0029] The processor has a built-in machine learning-based optimization algorithm to analyze the state of the damped wall based on sensor data and output optimization instructions.

[0030] The control module, according to the instructions output by the processor, controls the electro-hydraulic servo valve of the lifting hydraulic cylinder to adjust the viscosity depth, and is connected to the viscous fluid supply system to dynamically adjust the viscosity ratio of the viscous fluid.

[0031] Through the above technical solutions, the intelligent control system can identify seismic excitation characteristics and structural response status in real time based on multi-source sensor data and machine learning algorithms, and dynamically adjust the viscosity depth and viscosity of the damping wall, so that the damping wall can maintain optimal energy dissipation performance under seismic action of different intensities and different frequencies, thereby improving the seismic reliability and self-adaptive capability of the structure.

[0032] Preferably, the optimization algorithm built into the processor is a hybrid control algorithm based on the fusion of long short-term memory networks and model predictive control. The core calculation formula of this algorithm includes:

[0033] Step 1: Define the system state vector: ,in:

[0034] Let be the system state vector. For the displacement of the damping body. For the velocity of the damped body, For the damped body acceleration, Temperature of the viscous fluid. The average pressure on the surface of the viscous plate. For the strain of the wall pier.

[0035] Step 2: Define the control command vector: ,in:

[0036] To control the command vector, This is a viscosity depth adjustment command. This is a command to adjust the viscosity of viscous fluids.

[0037] Step 3, LSTM state prediction model: ,in:

[0038] To predict the state, For the prediction function, In hidden state, In cellular state, These are network parameters.

[0039] Step 4: Model Predictive Control Objective Function: ,in:

[0040] : Future control sequence to be optimized : Optimize the objective function, Ideal reference conditions State error, The weight matrix of the control input is a positive definite diagonal matrix used to balance tracking accuracy and control cost. The maximum permissible operating temperature for viscous fluids. : Indicates the current moment Future predictions based on LSTM network models and currently known information The system state vector at time t. Temperature over-limit penalty coefficient, : Indicates the current moment For the future The control command vector planned for each control cycle : This represents the predicted temperature of the viscous fluid at the current moment, for the next control cycle.

[0041] Step 5, Viscosity Dynamic Adjustment Formula: ,in, This is the viscosity compensation amount based on the constitutive model. For output based on the rule base, The root mean square of the acceleration, To estimate the magnitude, , These are the weighting coefficients.

[0042] The optimization algorithm, based on the above technical solution, includes real-time acquisition of sensor data, filtering and normalization processing, and formation of the current state vector. ,Will Inputting historical control inputs into an LSTM prediction model, and rolling forecasting of the future. Step state Solving the objective function in the prediction time domain The problem of minimizing the optimal control sequence is solved to obtain the optimal control sequence. Take the first element of the sequence As the current control command, it is output to the lifting hydraulic cylinder and viscous fluid supply system. Based on the actual execution results and the new round of sensor data, the prediction and optimization process is continuously updated to achieve closed-loop adaptive control.

[0043] The beneficial effects of this invention are as follows:

[0044] 1. By setting an adaptive mechanism, the damper body achieves multi-directional adaptive fine-tuning and overload protection in viscous fluid. This mechanism converts sliding friction into rolling friction through a ball bearing structure, significantly reducing motion resistance and wear. It ensures the smoothness of motion and self-balancing ability of the damper body under complex multi-dimensional vibrations, effectively avoiding additional bending moments and stress concentrations caused by the mismatch between the direction of motion and the rigidity of the structure. At the same time, its built-in butterfly spring buffer module can absorb huge impact energy under extreme overload, and together with the limit groove, it provides mechanical hard limit, providing double protection for the damping wall. This greatly enhances the structure's impact resistance and durability under rare earthquakes and extends the service life of the device.

[0045] 2. By setting up a viscosity-enhancing mechanism, the energy dissipation capacity and efficiency of the damping wall are significantly improved. The composite design of 180-degree symmetrically distributed airfoil plates combined with a network of surface spherical and cylindrical grooves greatly increases the contact area with the viscous liquid in three-dimensional space. This structure can effectively break the laminar boundary layer and stimulate strong fluid disturbance and shear at the microscale, thereby generating higher viscous energy dissipation per unit area. The airfoil plates are rigidly connected to the support shaft of the adaptive mechanism, which can dynamically optimize the fluid angle of attack as the adaptive mechanism rotates, ensuring efficient entry into the flow field under various motion postures and maximizing the generation of damping force.

[0046] 3. By setting up an intelligent control system based on multi-source sensors, the working state of the damping wall is perceived in real time and the parameters are adaptively optimized. The built-in intelligent algorithm that integrates long short-term memory network and model predictive control can analyze the seismic excitation characteristics in real time and predict the future state of the system. Then, it can solve for the optimal viscosity depth and viscosity adjustment command of the viscous fluid in a rolling manner. This allows the damping wall to dynamically adjust to the optimal working point according to the intensity, spectrum and actual structural response of the vibration. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of an adaptive viscous damping wall based on machine learning proposed in this invention;

[0048] Figure 2 This is a three-dimensional view of the damping body structure of an adaptive viscous damping wall based on machine learning proposed in this invention.

[0049] Figure 3 This is a three-dimensional view of a channel steel box structure for an adaptive viscous damping wall based on machine learning proposed in this invention.

[0050] Figure 4 This is a three-dimensional view of a viscous plate structure for an adaptive viscous damping wall based on machine learning proposed in this invention.

[0051] Figure 5This is a three-dimensional view of the connecting top plate structure of an adaptive viscous damping wall based on machine learning proposed in this invention.

[0052] Figure 6 This is a three-dimensional view of a dust cover structure for an adaptive viscous damping wall based on machine learning proposed in this invention.

[0053] Figure 7 This is a three-dimensional view of a butterfly spring structure for an adaptive viscous damping wall based on machine learning proposed in this invention.

[0054] Figure 8 This is a three-dimensional view of the window structure of an adaptive viscous damping wall based on machine learning proposed in this invention.

[0055] Figure 9 This is a three-dimensional view of the outer tube structure of an adaptive viscous damping wall based on machine learning proposed in this invention.

[0056] Figure 10 This is a three-dimensional view of a ball cage structure based on machine learning for an adaptive viscous damping wall proposed in this invention.

[0057] Figure 11 This is a three-dimensional view of the inner spherical structure of an adaptive viscous damping wall based on machine learning proposed in this invention.

[0058] Figure 12 This is a three-dimensional view of a ball-and-groove structure for an adaptive viscous damping wall based on machine learning proposed in this invention.

[0059] Figure 13 This is a block diagram of an intelligent control system for an adaptive viscous damping wall based on machine learning, as proposed in this invention.

[0060] In the diagram: 1. Channel steel box; 11. Lifting hydraulic cylinder; 12. Buffer telescopic rod; 13. Temperature sensor; 2. Damper connecting wall pier; 21. Strain sensor; 3. Damping body; 31. Connecting top plate; 32. Adhesive plate; 33. Acceleration sensor; 34. Displacement sensor; 35. Pressure sensor; 4. Adaptive mechanism; 41. Through hole; 42. Outer tube; 43. Elliptical groove; 44. Inner sphere; 45. Support shaft; 46. Ball cage; 47. Groove; 48. Window hole; 49. Ball; 50. Dust cover; 51. Force ring; 52. Limiting groove; 53. Buffer groove; 54. Butterfly spring; 6. Adhesion-enhancing mechanism; 61. Wing plate; 62. Ball groove; 63. Cylindrical groove. Detailed Implementation

[0061] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0062] Reference Figures 1-13 An adaptive viscous damping wall based on machine learning includes a channel steel box 1 filled with viscous fluid, which is set between two damper connecting wall piers 2. A damping body 3 is vertically placed inside the channel steel box 1. The upper surface of the damping body 3 is fixedly connected to the lower surface of the upper damper connecting wall pier 2 through a pre-embedded part. An adaptive mechanism 4 is set through the interior of the damping body 3. Adhesion-enhancing mechanisms 6 are set at both ends of the adaptive mechanism 4. The channel steel box 1 is made of high-strength steel, which has good rigidity and corrosion resistance. The damping body 3 is made of metal and its surface can be rust-proofed.

[0063] To adapt to the target damping force, the viscosity depth needs to be adjusted. Therefore, lifting hydraulic cylinders 11 are symmetrically distributed and hinged on the lower surface of the channel steel box 1. The piston rod end of the lifting hydraulic cylinder 11 is hinged to the upper surface of the lower damper connecting wall pier 2 through another embedded part. Multiple buffer telescopic rods 12 that support the channel steel box 1 are fixedly installed on the lower surface of the channel steel box 1 through several other sets of embedded parts. The lower surface of the buffer telescopic rod 12 is also fixedly installed to the upper surface of the lower damper connecting wall pier 2 through corresponding embedded parts. The lifting hydraulic cylinders 11 and the buffer telescopic rods 12 work together to adjust the height of the channel steel box 1 in real time during an earthquake to adapt to different vibration intensities. At the same time, the buffer telescopic rods 12 provide elastic support to avoid rigid impact and extend the service life of the device.

[0064] To achieve the damping effect brought about by the viscous fluid flow, the damping body 3 consists of a transverse connecting top plate 31 and a vertical viscous plate 32 fixedly connected to the lower surface of the connecting top plate 31. The upper surface of the connecting top plate 31 is fixedly connected to the lower surface of the upper damper connecting wall pier 2 through embedded parts. There are flow gaps between the peripheral surface and the lower surface of the viscous plate 32 and the inner wall of the channel steel box 1. A flow gap of 3-5mm is left between the viscous plate 32 and the inner wall of the channel steel box 1, so that the viscous fluid can form a stable flow field during vibration, enhancing the energy dissipation effect. The connecting top plate 31 is rigidly connected to the upper wall pier, directly transmitting the load and with strong structural integrity.

[0065] The damping body 3 reciprocates along the length of the damper connecting wall pier 2 via viscous fluid. During the movement, the adaptive mechanism 4 maintains self-balance for the damping body 3.

[0066] To enable self-adaptation between the damping body 3 and the viscous fluid, the adaptive mechanism 4 includes through holes 41 arranged in a rectangular array on the surface of the viscous plate 32. An outer sleeve 42 is fixedly fitted onto the inner surface of the through holes 41. Elliptical grooves 43 are arranged in a ring array on the inner surface of the outer sleeve 42. An inner sphere 44 with an inner hole is disposed inside the outer sleeve 42. A support shaft 45 penetrating both sides of the viscous plate 32 is fixedly connected to the inner surface of the inner hole of the inner sphere 44. The outer sleeve 42 serves as a fixed component, and its outer wall is fixed to the through holes 41 in the viscous plate 32 by interference fit or welding. The inner sphere 44 serves as a movable component, and its inner hole is fixed to the support shaft 45 by key connection or press fit, making the support shaft 45 and the inner sphere 44 an integral unit. The two ends of the support shaft 45 extend out of the viscous plate 32 for connecting to the outside. The viscosity-enhancing mechanism 6, due to the assembly gap between the inner sphere 44 and the inner wall of the outer sleeve 42, and the fact that the inner sphere 44 is not completely locked within the outer sleeve 42, gives the joint two rotational degrees of freedom. When the damper 3 moves in the viscous fluid and is subjected to fluid resistance or lateral inertial force from different directions, the support shaft 45 and the inner sphere 44 assembly can undergo a slight spherical rotation relative to the fixed outer sleeve 42 and viscous plate 32 assembly. This slight rotational capability allows the viscosity-enhancing mechanism 6 connected to both ends of the support shaft 45 to make adaptive angle fine adjustments according to the direction of the flow field force, thereby more effectively cutting into the viscous fluid flow, optimizing the damping force generation efficiency, and avoiding additional bending moments and stress concentrations caused by the mismatch between the direction of movement and the structural rigidity, significantly improving the smoothness of movement and structural durability of the damper 3.

[0067] To achieve the force transmission path and motion constraints, the adaptive mechanism 4 further includes a ball cage 46 slidably connected to the inner surface of the outer sleeve 42. The inner surface of the ball cage 46 is in sliding contact with the outer surface of the inner sphere 44. The outer surface of the inner sphere 44 has grooves 47 arranged in a ring array, corresponding one-to-one with the elliptical grooves 43. The outer surface of the ball cage 46 has multiple windows 48 evenly distributed circumferentially. Each window 48 rollsably accommodates a ball 49. The elliptical grooves 43 and grooves 47 are arranged opposite to each other and compress the ball 49, causing the ball 49 to roll in contact on the inner surfaces of the elliptical grooves 43 and grooves 47. The force and motion between the inner sphere 44 and the outer sleeve 42 are no longer transmitted through direct sliding friction. Instead, through the intermediate medium of the ball 49, the ball 49 simultaneously contacts the groove 47 on the inner ball 44 and the elliptical groove 43 on the outer tube 42. The geometry of the elliptical groove 43 generates a pre-constraint force on the ball 49. When the inner ball 44 attempts to rotate relative to the outer tube 42, the ball 49 will roll purely under the guidance of the groove 47 and the elliptical groove 43. This mechanism of converting sliding friction into rolling friction greatly reduces the frictional resistance and wear inside the joint, ensuring the sensitivity and low hysteresis of adaptive motion. The role of the ball cage 46 is to ensure that all the balls 49 always maintain the correct relative position and spacing, prevent them from colliding with each other or falling off, and make the load evenly distributed among the multiple balls 49.

[0068] To achieve dynamic sealing of the adaptive mechanism 4, a conical dust cover 50 is fixedly connected to the outer cylindrical surface of the open end of the outer sleeve 42 by a metal clamp. The other end of the dust cover 50 is fastened to the outer surface of the support shaft 45 by another metal clamp. The dust cover 50 is made of thermoplastic rubber and forms a sealed flexible cavity between the outer sleeve 42 and the support shaft 45. Grease is placed inside the flexible cavity. The conical or corrugated shape of the dust cover 50 provides it with axial and radial expansion and contraction capabilities. When the support shaft... When the inner ball 44 rotates due to the movement of the support shaft 45, the relative position between the support shaft 45 and the outer sleeve 42 will change slightly. The flexible dust cover 50 can be stretched, compressed or bent accordingly, always maintaining a sealed state without generating excessive restraint or fatigue cracking. The grease in the flexible cavity provides long-term lubrication for the contact surfaces of the ball 49, groove 47 and ball cage 46, and isolates corrosive components that may seep in from the viscous liquid. At the same time, the grease itself has a certain viscosity, which can play an additional buffering and damping role for the small vibrations of the joint.

[0069] The internal components of the adaptive mechanism 4, such as the outer sleeve 42 and the inner ball 44, are made of alloy steel or wear-resistant cast iron, and can withstand high temperatures up to 200°C or higher. To prevent impact on the adaptive mechanism 4 under extreme overload conditions and reduce its service life, the adaptive mechanism 4 also includes a force-bearing ring 51 fixedly sleeved on the outer surface of the middle part of the outer sleeve 42. The adhesive plate 32 also has a limiting groove 52 communicating with the through hole 41. The inner surface of the limiting groove 52 is movably sleeved with the outer surface of the force-bearing ring 51. The adhesive plate 32 also has symmetrically distributed buffer grooves 53 inside. The outer end of the buffer groove 53 communicates with the through hole 41 and the limiting groove 52. A butterfly spring 54 is fixedly connected to one inner wall of the buffer groove 53. The free end of the butterfly spring 54 holds the force-bearing ring... One side of the outer surface of 51 is wrapped and fixedly connected. When encountering an extremely large vibration that far exceeds the design standard, the axial force acting on the support shaft 45 and the inner ball 44 assembly may be abnormally large. At this time, the huge axial force will be transmitted to the force ring 51 through the ball 49 and the outer sleeve 42. The force ring 51 will absorb a large amount of impact energy through the deformation process of the butterfly spring 54. The limiting groove 52 is a cavity on the viscous plate 32 with a larger diameter than the through hole 41, which allows the force ring 51 to have a small axial movement space in it. It is connected to the through hole 41 and the buffer groove 53. Therefore, even if the viscous liquid penetrates into the buffer groove 53 and comes into contact with the butterfly spring 54, its deformation process will force the viscous liquid to adhere to the surface of the butterfly spring 54, thereby playing an additional buffering and damping role.

[0070] By setting the adaptive mechanism 4, the damper 3 achieves multi-directional adaptive fine-tuning and overload protection in viscous fluid. This mechanism converts sliding friction into rolling friction through the ball bearing 49 structure, which significantly reduces motion resistance and wear, ensuring the smoothness of motion and self-balancing ability of the damper 3 under complex multi-dimensional vibrations. It effectively avoids additional bending moments and stress concentrations caused by the mismatch between the direction of motion and the rigidity of the structure. At the same time, its built-in butterfly spring 54 buffer module can absorb huge impact energy under extreme overload, and together with the limiting groove 52, it provides mechanical hard limiting, providing double protection for the damping wall, greatly enhancing the impact resistance and durability of the structure under rare earthquakes, and extending the service life of the device.

[0071] The adhesive-enhancing mechanism 6 is in full contact with the viscous liquid inside the channel steel box 1 and provides adhesive damping during vibration.

[0072] The adhesion-enhancing mechanism 6 includes wing plates 61 fixedly connected to the outer surfaces of the two free ends of the support shaft 45. The wing plates 61 of the adhesion-enhancing mechanism 6 are made of aluminum alloy or stainless steel, with smooth surfaces and shear resistance. In order to increase the lateral contact area and interaction strength with the viscous liquid, thereby significantly improving the damping performance, the two wing plates 61 are symmetrically distributed around the center point of the support shaft 45 at 180 degrees. The outer edges of the wing plates 61 are all rounded. The upper and lower surfaces of the wing plates 61 are both provided with regularly arranged surface features. The ball groove 62 is fixedly connected to adjacent ball grooves 62 by cylindrical grooves 63, thus forming a grid-like network of connected grooves on the upper and lower surfaces of the wingplate 61. The wingplate 61 is a smooth and full sheet-like structure, and its large surface area provides basic planar contact. The two wingplates 61 are arranged symmetrically at 180 degrees, ensuring that in the reciprocating motion of the damper 3, regardless of the direction of motion, at least one wingplate 61 can provide symmetrical and balanced damping force with a better angle of attack and viscous fluid action. The ball groove 62 serves as a local concave shape. The column groove 63 serves as a connecting channel. When the airfoil 61 moves in the viscous fluid, the viscous fluid flows over its surface. The network of interconnected grooves disrupts the stable development of the laminar boundary layer on the flat plate, forcing the fluid to enter, pass through, and flow out of the various spherical grooves 62 and cylindrical grooves 63. This process continuously generates microscale eddies, deflection, and reattachment, greatly increasing the shear deformation rate and velocity gradient inside the fluid. The rotation function of the adaptive mechanism 4 allows the airfoil 61 to adaptively adjust its angle relative to the incoming flow according to the instantaneous flow field direction. This dynamic adjustment ensures that the airfoil 61, based on the reinforcement effect of the network grooves, can always enter the flow field with a better hydrodynamic attitude, maximizing its efficiency in pushing the viscous fluid and generating shear force, and avoiding performance loss or vibration coupling problems that may be caused by a fixed angle. The airfoil 61 is preferably made of high-strength aluminum alloy or austenitic stainless steel, meeting the working temperature range of -30℃ to 150℃, to adapt to the temperature rise that the viscous fluid may generate due to repeated shearing during the earthquake, as well as the temperature changes of the external environment where the building is located, ensuring structural stability and fatigue life.

[0073] By setting up the viscosity-enhancing mechanism 6, the energy dissipation capacity and efficiency of the damping wall are significantly improved. The composite design of the 180-degree symmetrically distributed airfoil 61 combined with the network of surface spherical grooves 62 and cylindrical grooves 63 greatly increases the contact area with the viscous liquid in three-dimensional space. This structure can effectively break the laminar boundary layer and stimulate strong fluid disturbance and shear at the microscale, thereby generating higher viscous energy dissipation per unit area. The airfoil 61 is rigidly connected to the support shaft 45 of the adaptive mechanism 4, and its fluid angle of attack can be dynamically optimized with the rotation of the adaptive mechanism 4 to ensure efficient entry into the flow field under various motion attitudes and maximize the generation of damping force.

[0074] The viscous damping wall includes a sensor system and an intelligent control system. The sensor system includes:

[0075] The displacement sensor 34 and acceleration sensor 33 installed on the damper 3 are used to monitor the displacement, velocity and acceleration of the damper 3 in the viscous liquid in real time.

[0076] Temperature sensor 13 is installed on the inner wall of channel steel box 1.

[0077] Pressure sensor 35 is mounted on the surface of the adhesive plate 32.

[0078] Strain sensor 21 is installed on the damper connecting wall pier 2.

[0079] The intelligent control system includes:

[0080] The data acquisition module connects to all sensors and is used to collect sensor data in real time.

[0081] The processor has a built-in machine learning-based optimization algorithm to analyze the state of the damped wall based on sensor data and output optimization instructions.

[0082] The control module, according to the instructions output by the processor, controls the electro-hydraulic servo valve of the lifting hydraulic cylinder 11 to adjust the viscosity depth, and connects to the viscous fluid supply system to dynamically adjust the viscosity ratio of the viscous fluid. The intelligent control system can identify the seismic excitation characteristics and structural response status in real time based on multi-source sensor data and through machine learning algorithms, and dynamically adjust the viscosity depth and viscosity of the damping wall, so that the damping wall can maintain the optimal energy dissipation performance under seismic action of different intensities and different frequencies, thereby improving the seismic reliability and self-adaptive ability of the structure.

[0083] The processor's built-in optimization algorithm is a hybrid control algorithm based on the fusion of long short-term memory networks and model predictive control. The core calculation formulas of this algorithm include:

[0084] Step 1: Define the system state vector: ,in:

[0085] Let be the system state vector. The displacement of the damping body is 3. For the damped body, 3 velocity, For the damped body 3 acceleration, Temperature of the viscous fluid. The average pressure on the surface of the viscous plate 32. For the strain of the wall pier.

[0086] Step 2: Define the control command vector: ,in:

[0087] To control the command vector, This is a viscosity depth adjustment command. This is a command to adjust the viscosity of viscous fluids.

[0088] Step 3, LSTM state prediction model: ,in:

[0089] To predict the state, For the prediction function, In hidden state, In cellular state, These are network parameters.

[0090] Step 4: Model Predictive Control Objective Function: ,in:

[0091] : Future control sequence to be optimized : Optimize the objective function, Ideal reference conditions State error, The weight matrix of the control input is a positive definite diagonal matrix used to balance tracking accuracy and control cost. The maximum permissible operating temperature for viscous fluids. : Indicates the current moment Future predictions based on LSTM network models and currently known information The system state vector at time t. Temperature over-limit penalty coefficient, : Indicates the current moment For the future The control command vector planned for each control cycle : This represents the predicted temperature of the viscous fluid at the current moment, for the next control cycle.

[0092] Step 5, Viscosity Dynamic Adjustment Formula: ,in, This is the viscosity compensation amount based on the constitutive model. For output based on the rule base, The root mean square of the acceleration, To estimate the magnitude, , These are the weighting coefficients.

[0093] The optimization algorithm's execution steps include real-time acquisition of sensor data, filtering and normalization processing, and formation of the current state vector. ,Will Inputting historical control inputs into an LSTM prediction model, and rolling forecasting of the future. Step state Solving the objective function in the prediction time domain The problem of minimizing the optimal control sequence is solved to obtain the optimal control sequence. Take the first element of the sequence As the current control command, it is output to the lifting hydraulic cylinder 11 and the viscous fluid supply system. Based on the actual execution results and the new round of sensor data, the prediction and optimization process is continuously updated to achieve closed-loop adaptive control.

[0094] By setting up an intelligent control system based on multi-source sensors, the working state of the damping wall is perceived in real time and the parameters are optimized adaptively. The built-in intelligent algorithm that integrates long short-term memory network and model predictive control can analyze the seismic excitation characteristics in real time and predict the future state of the system. Then, it can solve for the optimal viscosity depth and viscosity adjustment command of the viscous fluid in a rolling manner. This allows the damping wall to dynamically adjust to the optimal working point according to the intensity, spectrum and actual response of the vibration.

[0095] Working principle: In a specific embodiment of the present invention, when the building structure is stimulated by external factors, relative motion will occur between the upper and lower dampers connected to the wall pier 2. This motion forces the damper 3 to reciprocate vertically within the channel steel box 1 filled with viscous liquid. When the damper 3 moves, the viscous liquid between its viscous plate 32 and the inner wall of the channel steel box 1 is subjected to strong shearing, which converts the kinetic energy of the structure into heat energy, thereby achieving energy dissipation and vibration reduction.

[0096] When the damper 3 moves in the viscous fluid, it will be subjected to multidimensional and non-uniform resistance from the fluid. At this time, the adaptive mechanism 4 embedded in the viscous plate 32 is activated. When subjected to lateral fluid force or inertial force, the inner ball 44 undergoes a small spherical rotation in the outer tube 42, that is, the ball 49 rolls in the elliptical groove 43 and groove 47, which greatly reduces friction. This rotation enables the entire support shaft 45 and the airfoil 61 at both ends to make real-time angle fine adjustments in accordance with the flow field direction. The effect is that the damper 3 is no longer a rigid whole, but has a flexible joint, which can adaptively move with the flow, thereby significantly reducing the additional bending moment, stress concentration and unnecessary lateral resistance during the movement, making the movement smoother and improving durability.

[0097] When encountering an extreme vibration far exceeding the design standard, the huge axial impact force is transmitted to the outer sleeve 42 through the support shaft 45, inner ball 44, and ball 49, and then acts on the force ring 51 on it. The force ring 51 compresses the butterfly spring 54 on one side. The butterfly spring 54 absorbs and dissipates the huge impact energy through its own large deformation. The limiting groove 52 provides space for the movement of the force ring 51 and the final mechanical hard limit to prevent the structure from being crushed.

[0098] The wing plates 61, fixed at both ends of the support shaft 45, directly participate in shearing the viscous fluid with their large surface area. The unique spherical grooves 62 and cylindrical grooves 63 on their surfaces play a key role: the spherical grooves 62 and cylindrical grooves 63 make the effective shearing area far exceed the projected area of ​​the flat plate. During movement, the spherical grooves 62 and cylindrical grooves 63 cause the viscous fluid to generate complex micro-scale eddies and re-attachment, which greatly increases the local shear rate and thus generates a greater damping force at the same movement speed.

[0099] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An adaptive viscous damping wall based on machine learning, comprising a channel steel box (1) internally loaded with viscous fluid, characterized in that: The channel steel box (1) is set between two damper connecting wall piers (2). A damping body (3) is vertically placed inside the channel steel box (1). The upper surface of the damping body (3) is fixedly connected to the lower surface of the upper damper connecting wall pier (2) through a pre-embedded part. An adaptive mechanism (4) is set through the interior of the damping body (3). Adhesion-enhancing mechanisms (6) are set at both ends of the adaptive mechanism (4). The damping body (3) reciprocates along the length of the damper connecting wall pier (2) via viscous fluid. During the movement, the adaptive mechanism (4) maintains self-balance for the damping body (3). The adhesive-enhancing mechanism (6) is in full contact with the viscous liquid inside the channel steel box (1) and provides adhesive damping during vibration.

2. The adaptive viscous damping wall based on machine learning according to claim 1, characterized in that: The lower surface of the channel steel box (1) is symmetrically distributed and hinged with lifting hydraulic cylinders (11). The piston rod end of the lifting hydraulic cylinder (11) is hinged to the upper surface of the lower damper connecting wall pier (2) through another embedded part. The lower surface of the channel steel box (1) is also fixedly installed with multiple buffer telescopic rods (12) to support the channel steel box (1) through several other sets of embedded parts. The lower surface of the buffer telescopic rod (12) is also fixedly installed to the upper surface of the lower damper connecting wall pier (2) through corresponding embedded parts.

3. The adaptive viscous damping wall based on machine learning according to claim 2, characterized in that: The damping body (3) consists of a transverse connecting top plate (31) and a vertical adhesive plate (32) fixedly connected to the lower surface of the connecting top plate (31). The upper surface of the connecting top plate (31) is fixedly connected to the lower surface of the upper damper connecting wall pier (2) through embedded parts. The peripheral surface and lower surface of the adhesive plate (32) have flow gaps with the inner wall of the channel steel box (1).

4. The adaptive viscous damping wall based on machine learning according to claim 3, characterized in that: The adaptive mechanism (4) includes through holes (41) arranged in a rectangular array on the surface of the adhesive plate (32). An outer sleeve (42) is fixedly sleeved on the inner surface of the through holes (41). Elliptical grooves (43) are arranged in a ring array on the inner surface of the outer sleeve (42). An inner sphere (44) with an inner hole is provided inside the outer sleeve (42). A support shaft (45) that passes through both sides of the adhesive plate (32) is fixedly connected to the inner surface of the inner hole of the inner sphere (44).

5. The adaptive viscous damping wall based on machine learning according to claim 4, characterized in that: The adaptive mechanism (4) further includes a ball cage (46) slidably connected to the inner surface of the outer sleeve (42). The inner surface of the ball cage (46) is in sliding contact with the outer surface of the inner sphere (44). The outer surface of the inner sphere (44) is provided with grooves (47) in a ring array that correspond one-to-one with the elliptical groove (43). The outer surface of the ball cage (46) is provided with a plurality of windows (48) evenly distributed along the circumference. Each window (48) slidably accommodates a ball (49). The elliptical groove (43) is arranged opposite to the groove (47) and squeezes the ball (49) so that the ball (49) rolls in contact with the inner surface of the elliptical groove (43) and the groove (47).

6. The adaptive viscous damping wall based on machine learning according to claim 5, characterized in that: A conical dust cover (50) is fixedly connected to the outer cylindrical surface of the opening end of the outer sleeve (42) by a metal clamp. The other end of the dust cover (50) is fastened to the outer surface of the support shaft (45) by another metal clamp. The dust cover (50) is made of thermoplastic rubber and forms a closed flexible cavity between the outer sleeve (42) and the support shaft (45). Lubricating grease is provided in the flexible cavity.

7. The adaptive viscous damping wall based on machine learning according to claim 6, characterized in that: The adaptive mechanism (4) further includes a force ring (51) fixedly sleeved on the outer surface of the middle part of the outer sleeve (42). The adhesive plate (32) also has a limiting groove (52) communicating with the through hole (41). The inner surface of the limiting groove (52) is movably sleeved with the outer surface of the force ring (51). The adhesive plate (32) also has a buffer groove (53) symmetrically distributed inside. The outer end of the buffer groove (53) is connected to the through hole (41) and the limiting groove (52). A butterfly spring (54) is fixedly connected to one side of the inner wall of the buffer groove (53). The free end of the butterfly spring (54) wraps around and fixes one side of the outer surface of the force ring (51).

8. The adaptive viscous damping wall based on machine learning according to claim 7, characterized in that: The adhesion enhancement mechanism (6) includes wing plates (61) that are fixedly connected to the outer surfaces of the two free ends of the support shaft (45). The two wing plates (61) are symmetrically distributed around the center point of the support shaft (45) at 180 degrees. The outer edges of the wing plates (61) are all rounded. The upper and lower surfaces of the wing plates (61) are provided with regularly arranged ball grooves (62). The adjacent ball grooves (62) are fixedly connected by cylindrical grooves (63), thereby forming a grid-like network of connected grooves on the upper and lower surfaces of the wing plates (61).

9. The adaptive viscous damping wall based on machine learning according to claim 8, characterized in that: The viscous damping wall includes a sensor system and an intelligent control system, wherein the sensor system includes: The displacement sensor (34) and acceleration sensor (33) installed on the damper (3) are used to monitor the displacement, velocity and acceleration of the damper (3) in the viscous liquid in real time. Temperature sensor (13) installed on the inner wall of the channel steel box (1); Pressure sensor (35) mounted on the surface of the viscous plate (32); Strain sensor (21) installed on the damper connecting wall pier (2); The intelligent control system includes: The data acquisition module connects to all sensors and is used to acquire sensor data in real time. The processor has a built-in machine learning-based optimization algorithm to analyze the state of the damped wall based on sensor data and output optimization instructions. The control module controls the electro-hydraulic servo valve of the lifting hydraulic cylinder (11) to adjust the viscosity depth according to the instructions output by the processor, and connects to the viscosity supply system to dynamically adjust the viscosity ratio of the viscosity liquid.

10. The adaptive viscous damping wall based on machine learning according to claim 9, characterized in that: The processor's built-in optimization algorithm is a hybrid control algorithm based on the fusion of long short-term memory networks and model predictive control. The core calculation formula of this algorithm includes: Step 1: Define the system state vector: ,in: Let be the system state vector. For the displacement of the damping body. For the velocity of the damped body, For the damped body acceleration, Temperature of the viscous fluid. The average pressure on the surface of the viscous plate. For the strain of the wall pier; Step 2: Define the control command vector: ,in: To control the command vector, This is a viscosity depth adjustment command. This is a command to adjust the viscosity of viscous fluids. Step 3, LSTM state prediction model: ,in: To predict the state, For the prediction function, In hidden state, In cellular state, For network parameters; Step 4: Model Predictive Control Objective Function: ,in: : Future control sequence to be optimized : Optimize the objective function, Ideal reference conditions State error, The weight matrix of the control input is a positive definite diagonal matrix used to balance tracking accuracy and control cost. The maximum permissible operating temperature for viscous fluids. : Indicates the current moment Future predictions based on LSTM network models and currently known information The system state vector at time t. Temperature over-limit penalty coefficient, : Indicates the current moment For the future The control command vector planned for each control cycle : Indicates the predicted temperature of the viscous fluid at the current moment for the next control cycle; Step 5, Viscosity Dynamic Adjustment Formula: ,in, This is the viscosity compensation amount based on the constitutive model. For output based on the rule base, The root mean square of the acceleration, To estimate the magnitude, , These are the weighting coefficients.

Images