Active vibration reduction and intelligent airflow coordination control system for clean room

Abstract

The application discloses an active damping and intelligent airflow coordination control system for a clean room, and relates to the technical field of purification engineering. Technical scheme highlights of the application are as follows: the system comprises an equipment bearing platform, an active damping module, a micro environment management module, a sensor network and a central coordination controller, the active damping module is arranged below the equipment bearing platform and is used for offsetting vibration transmitted to the equipment bearing platform, the micro environment management module comprises a core working area surrounding equipment and a support frame decoupled from vibration of the equipment bearing platform, and the micro environment management module is used for providing directional clean airflow to the core working area and trapping pollutants. Through arrangement of various modules and coordination control between the modules, the application not only guarantees a clean environment of the core working area, but also avoids introducing new vibration disturbance to the equipment bearing platform.

Description

Technical Field

[0001] This invention relates to the field of cleanroom engineering technology, and more specifically, to an active vibration reduction and intelligent airflow coordinated control system for cleanrooms. Background Technology

[0002] In modern industrial cleanrooms, especially in high-end manufacturing fields such as semiconductors and precision optics, in order to meet the extreme requirements of core processes for the microenvironment, active vibration isolation platforms are usually deployed inside the cleanroom to place equipment on the platform to isolate ground vibrations, while air cleanliness is maintained by air supply systems such as overhead fan filter units.

[0003] However, the different subsystems are physically separated and control logic is independent of each other. This makes it impossible for them to achieve unified prediction and coordinated response when dealing with the complex disturbances generated by the dynamic operation of the process equipment itself. The system can only perform passive and isolated feedback compensation after vibration, thermal disturbance or contamination actually occurs and is detected. This makes the core process area of ​​the cleanroom always have periodic instantaneous microenvironment instability.

[0004] Therefore, a new solution is needed to address this problem. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an active vibration reduction and intelligent airflow coordinated control system for clean rooms.

[0006] The above-mentioned technical objective of the present invention is achieved through the following technical solution: an active vibration reduction and intelligent airflow coordinated control system for cleanrooms, including an equipment support platform, an active vibration reduction module, a microenvironment management module, a sensor network and a central coordinated controller, wherein the active vibration reduction module is disposed below the equipment support platform and is used to counteract the vibration transmitted to the equipment support platform; The microenvironment management module includes a core working area surrounding the equipment and a support frame decoupled from the vibration of the equipment's carrying platform. The microenvironment management module is used to provide directional clean airflow to the core working area and capture pollutants. The sensor network is used to detect the vibration of the equipment carrying platform and the environmental state parameters of the core working area. The environmental state parameters include at least temperature field parameters and pollutant concentration parameters, so as to obtain real-time monitoring data. The central coordinating controller is connected to the active vibration reduction module, the micro-environment management module, the sensor network, and the control system signals of the equipment. The central coordinating controller is configured as follows: Store and recall preset microenvironment expected targets, including vibration limits, temperature fluctuation limits, and particle concentration limits; Receive real-time motion instructions from the control system. The real-time motion instructions include motion types and motion parameters that describe the actions that the device is about to perform. The real-time motion instructions are automation instructions within the control system. Based on its internally stored equipment disturbance feature model, the system predicts the multi-physics disturbance information that will be triggered according to real-time motion commands. The multi-physics disturbance information includes at least the predicted vibration spectrum and the predicted heat generation location information. Based on the predicted vibration spectrum, a first feedforward control command is generated and sent to the active vibration reduction module, so that the active vibration reduction module generates a canceling force with the same amplitude and opposite phase as the predicted vibration spectrum. Based on the predicted heat generation location information, a second feedforward control command is generated and sent to the microenvironment management module to adjust the clean airflow parameters at the corresponding location. Receive real-time monitoring data from the sensor network, and calculate the deviation between the real-time monitoring data and the expected target; Feedback control commands are generated based on the deviation between the real-time monitoring data and the expected target, and simultaneously sent to the active vibration reduction module and the micro-environment management module to coordinately adjust the output force of the active vibration reduction module and the airflow parameters of the micro-environment management module until the deviation converges to a preset threshold range.

[0007] The present invention is further configured such that: the active vibration reduction module includes a plurality of active vibration isolation units, the equipment support platform is disposed on top of the plurality of active vibration isolation units, and the active vibration isolation unit includes an air spring, a voice coil motor actuator and a high-precision accelerometer.

[0008] The invention is further configured such that: the stator portion of the voice coil motor actuator is fixedly connected to the top plate of the air spring; the mover portion of the voice coil motor constitutes the top output end of the active vibration isolation unit; and the high-precision accelerometer is fixedly connected to the mover portion of the voice coil motor actuator for measuring the acceleration of the output end.

[0009] The present invention is further configured such that: a plurality of fan filter units are provided on the support frame above the core working area, the fan filter units are used to provide clean airflow vertically downward, and at least one adjustable directional airflow arm is provided around the plurality of fan filter units, and a composite nozzle is provided at the end of the adjustable directional airflow arm.

[0010] The present invention is further configured such that: the interior of the composite nozzle is divided into an air supply duct and an air exhaust duct, the air supply duct is used to deliver airflow to the predicted heat generation location to form an airflow barrier, and the air exhaust duct is used to form a negative pressure trapping zone around the airflow barrier.

[0011] The present invention is further configured such that: the sensor network includes a vibration sensor, a thermal imager and a particle counter; the vibration sensor is configured as a high-precision accelerometer and a vibration probe; the thermal imager is fixedly connected to the support frame and its field of view covers the core working area; and the isodynamic sampling port of the particle counter is integrated into the composite nozzle of the adjustable directional airflow arm.

[0012] The present invention is further configured such that the central coordinating controller is also configured to superimpose the first feedforward control command, the second feedforward control command, and the feedback control command in the time domain to form a final output integrated control command, wherein the first feedforward control command and the second feedforward control command are used to establish a reference value for the system response, and the feedback control command is used to compensate for the residual errors of the first feedforward control command and the second feedforward control command and external sudden disturbances.

[0013] The present invention is further configured as follows: a method for active vibration reduction and intelligent airflow coordinated control of a cleanroom for implementing the system, characterized in that: the method includes: S1. Obtain real-time motion commands from the equipment control system; S2. Based on the equipment disturbance feature model pre-stored in the central coordinating controller, predict the multi-physics disturbance information that will be triggered according to real-time motion commands. The multi-physics disturbance information includes at least the predicted vibration spectrum and the predicted heat generation location information. S3. Generate and execute the first feedforward control command to the active vibration reduction module to actively generate a counteracting force based on the predicted vibration spectrum; S4. Generate and execute a second feedforward control command to the microenvironment management module to adjust the clean airflow parameters at the corresponding location based on the predicted heat generation location information; S5. Acquire real-time monitoring data from the sensor network; S6. Generate feedback control commands based on the deviation between real-time monitoring data and expected targets, so as to coordinate the output force of the active vibration reduction module and the airflow parameters of the microenvironment management module. Step S4 further includes: S4.1 Control the adjustable directional airflow arm to move to the coordinates corresponding to the predicted heat generation location information; S4.2 Adjust the supply air velocity and exhaust air volume of the composite nozzle at the end of the adjustable directional airflow arm to suppress the predicted thermal plume and capture the emitted pollutants. In summary, the present invention has the following beneficial effects: by setting up vibration decoupling of the support frame and the layout of the fan filter unit and the adjustable directional airflow arm on the support frame, the local enhanced airflow and the background clean airflow work in a layered and coordinated manner, which not only ensures the clean environment of the core working area, but also avoids introducing new vibration disturbances to the equipment carrying platform. By setting up the air supply duct and exhaust duct, directional airflow protection is formed at the equipment opening or heat generation position, realizing the source capture of pollutants and the active suppression of hot plumes. Attached Figure Description

[0014] Figure 1 This is a control system framework diagram of the device in this invention; Figure 2 This is a flowchart of the collaborative control method in this invention. Detailed Implementation

[0015] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0016] An active vibration damping and intelligent airflow coordinated control system for cleanrooms, such as Figure 1 and Figure 2As shown, this system includes an equipment support platform, an active vibration damping module, a microenvironment management module, a sensor network, and a central collaborative controller. The equipment support platform can be made of natural granite with a low coefficient of thermal expansion and high damping. Its upper surface is used to install precision process equipment, such as a deep ultraviolet lithography machine. The lower surface of the equipment support platform is provided with mounting interfaces corresponding to the active vibration isolation units. The active vibration damping module includes at least four active vibration isolation units, which are arranged in a rectangular array. Each active vibration isolation unit includes an air spring, a voice coil motor actuator, and a high-precision accelerometer. The air spring is at the bottom of the active vibration isolation unit and is in direct contact with the cleanroom floor. The air spring is filled with high-pressure nitrogen, and the air pressure is regulated by an external servo valve. Therefore, it can withstand approximately 95% of the static weight of the equipment and the equipment support platform and provide low-frequency passive vibration isolation in both horizontal and vertical directions. The stator of the voice coil motor actuator is constructed with high-strength... The top plate of the air spring is bolted with a moving part that is a power output shaft that can move along the vertical axis, thus forming the top power output end of the active vibration isolation unit. The top power output end is fixedly connected to the bottom surface of the equipment support platform through a flange and high-preload bolts. A high-precision accelerometer is bonded to the surface of the moving part of the voice coil motor actuator and keeps synchronized with the moving part of the top power output end. The accelerometer is used to measure the absolute acceleration signal of the power output end in real time and feed it back to the central coordinating controller. Several active vibration isolation units are supported and connected to the same equipment support platform through their top power output ends. The equipment support platform and each power output end use positioning pins to achieve high positioning accuracy and are locked with several sets of high-preload hexagonal socket bolts to ensure that the connection interface has high dynamic stiffness. This allows the control force generated by the voice coil motor actuator to be transmitted to the equipment support platform and the equipment above without loss or delay, improving the response speed of vibration control.

[0017] The microenvironment management module includes an independent support frame, along with fan filter units and adjustable directional airflow arms mounted on the support frame. The support frame is constructed from high-strength aluminum alloy profiles, with four uprights fixed to the bottom of the cleanroom. It is completely independent of the equipment support platform and active vibration damping module. There is no rigid contact between the equipment frame and the equipment support platform; they are only sealed together via flexible corrugated pipes, achieving complete vibration decoupling and effectively preventing secondary disturbances introduced by the microenvironment management module's own movement. On the top beam of the support frame, four fan filter units arranged in a rectangular array are installed, facing the core working area of ​​the equipment below. Each fan filter unit consists of an EC fan, a high-efficiency ULPA filter, and a flow guide plate, allowing independent airflow adjustment. This provides a stable, uniform, and vertically downward background clean airflow to the core working area, maintaining a basic clean environment. On the side uprights of the support frame, below the fan filter units, are two adjustable directional airflow arms. These airflow arms are six-degree-of-freedom tandem articulated robotic arms, with their bases fixed to the support frame and their ends... The actuator is a composite nozzle, whose interior is completely divided into two independent flow channels by a longitudinal baffle: an air supply duct and an exhaust duct. The air supply duct is connected to an independent clean compressed air source, and its outlet is a 0.5mm wide slit, capable of ejecting highly uniform laminar airflow to form an air knife-like barrier. This barrier has extremely high flow stability and anti-disturbance stiffness, forming a gaseous isolation wall in front of equipment openings or heat-generating locations. This effectively blocks air exchange between the core working area and the external environment, preventing external particles from diffusing and intruding with turbulent flow, while simultaneously protecting the equipment. The flow of pollutants emitted from within provides a predetermined direction. The exhaust duct surrounds the outside of the supply duct, and the inlet is a gradually expanding horn-shaped opening connected to the central vacuum exhaust system. This is used to create a stable negative pressure trapping zone with a large coverage area and low flow velocity gradient around the barrier. This trapping zone and the purging direction of the barrier form spatial complementarity and flow field coupling, allowing pollutants blown away from the core working area by the air knife to immediately enter the negative pressure zone and be instantly extracted, avoiding secondary diffusion and resuspension of pollutants. The structural design reduces the dependence of the core working area on the overall air exchange rate.

[0018] The sensor network includes vibration sensors, a thermal imager, and a particle counter. The vibration sensors include high-precision accelerometers and non-contact vibration probes. These two devices perform different measurement tasks. The high-precision accelerometer is mounted on the mover surface of the voice coil motor actuator in each active vibration isolation unit. This position coincides with the application point of the vibration control force and is in the same rigid body motion mode as the equipment mounting reference plane. Measuring acceleration at this location allows for zero-phase-lag capture of the overall motion state of the platform, a necessary condition for achieving high-bandwidth closed-loop feedback control. The high-precision accelerometer monitors the residual vibration acceleration, velocity, and displacement of the equipment's bearing platform in six degrees of freedom. The non-contact... The contact vibration probe can be a laser Doppler vibrometer or an eddy current displacement sensor. The vibration probe is fixed on a support frame. The laser beam or probe is an extremely sensitive component of the non-contact alignment equipment. This component includes, but is not limited to, the lower end face of the projection lens of a lithography machine, which is the end of the imaging optical path. Micron-level vibrations can cause overlay accuracy failure. The electron tube pole shoe of an electron beam lithography machine, where electromagnetic fields are coupled with mechanical vibrations, will directly affect beam spot positioning. The edge or upper surface of the workpiece stage, which is the direct bearing surface of the moving parts, can reflect the most accurate process position stability. The optical window or sensor bracket of the alignment system, where vibrations will directly cause a decrease in the signal-to-noise ratio of the alignment signal.

[0019] The thermal imager is fixed on the top beam or side cantilever of the support frame. Its lens optical axis is vertically downward or tilted to cover the core working area of ​​the equipment. The specific coverage area includes the workpiece stage movement area, which is the main heat source of the stepping scanning motion. The linear motor coil and guide rail friction points are all located in this area. The wafer bearing surface, where temperature uniformity directly determines the consistency of critical process dimensions, the reaction chamber or deposition source opening, where local high-temperature gas may escape, and specific heat-generating windows of the equipment casing, such as cooling water interfaces and near high-frequency power modules. The thermal imager is a non-contact, high-frame-rate temperature field sensing device. Deploying it on the support frame can obtain a panoramic thermal map of the entire core area. Because it is decoupled from the vibration of the equipment bearing platform, it avoids interference with the imaging quality caused by its own vibration and will not affect the thermal balance of the equipment. This deployment method enables the system to identify the starting position and diffusion path of the thermal plume in real time, providing precise spatial coordinate commands and dynamic wind speed compensation basis for the adjustable directional airflow arm.

[0020] The isodynamic sampling port of the particle counter is integrated inside the composite nozzle at the end of the adjustable directional airflow arm, specifically located in the transition area between the supply and exhaust ducts. For special workstations where the airflow arm cannot be deployed, the sampling port of the particle counter can be independently fixed on the cantilever of the support frame, adjacent to key openings of the equipment, such as wafer transfer ports, observation windows, and valve interfaces, with a distance controlled between 5-20mm. By placing the sampling port of the particle counter inside the composite nozzle, the particle counter can directly measure the airflow sample that has just been isolated by the barrier and is about to be captured by negative pressure. This is the location with the highest, most drastic, and earliest detectable pollutant concentration in the entire cleanroom air, enabling nanosecond-level early warning before pollutants diffuse into the core process area. The isodynamic design ensures that the sampling port is oriented in the same direction as the airflow, and the sampling flow rate matches the mainstream flow rate, ensuring that the particle size distribution is not distorted and the counting concentration is without deviation. Because the port moves synchronously with the adjustable directional airflow arm, the system can perform real-time monitoring of each predicted heat and pollution generation location, improving the sensing capability.

[0021] The central coordinating controller is an industrial control computer based on an Intel Xeon processor and a real-time operating system. It connects via EtherCAT bus to the voice coil motor servo driver of the active vibration isolation unit, the joint motor driver of the adjustable directional airflow arm, the frequency converter of the fan filter unit, the vacuum exhaust valve, and all sensor acquisition modules. Simultaneously, the controller establishes a real-time data link with the lithography machine's main control system through the SECS / GEM communication protocol. The central coordinating controller internally stores a pre-calibrated equipment disturbance characteristic model. This model is a database of mapping relationships, recording the correspondence between each typical motion command of the lithography machine, such as a 10mm step in the X direction of the worktable and the opening of the wafer exchange door, and the resulting vibration spectrum characteristics, motor heating location and power curve, and particle emission risk. The central coordinating controller also internally stores microenvironmental expectations. These target values ​​are derived from installation specifications, process sensitivity, and cleanroom design standards provided by the equipment supplier, including vibration limits, temperature fluctuation limits, and particle concentration limits. The core control logic of the controller is as follows: Feedforward prediction: Real-time capture of the next motion command that the lithography machine is about to execute, query the disturbance feature model, and immediately calculate the required counteracting force spectrum and the coordinates of the expected heat generation location.

[0022] Coordination command issuance: On the one hand, a pre-generated reverse force current waveform command is sent to the servo driver of the voice coil motor actuator of the four active vibration isolation units; on the other hand, a motion command is sent to the adjustable directional airflow arm, causing its composite nozzle to move quickly to the side of the predicted heat generation position, and simultaneously adjusting the supply air speed and exhaust air volume to the preset values, while slightly increasing the fan speed of the corresponding fan filter unit above that area.

[0023] Feedback correction: Real-time data from all accelerometers, laser vibrometers, thermal imagers, and particle counters are collected at a rate of 20,000 times per second. The deviation from the target values ​​(such as residual vibration less than 1 nm, temperature fluctuation less than 0.01℃, and zero particle concentration detection) is calculated, and feedback compensation commands are generated through a multivariable PID algorithm to fine-tune the output force of each voice coil motor and the airflow parameters of each airflow actuator in real time.

[0024] Solution and mapping: The built-in six-degree-of-freedom decoupling algorithm is used to accurately decompose the total control torque calculated based on the overall motion mode of the platform into component force commands for the voice coil motors of each axis of the four vibration isolation units, ensuring precise and coordinated control of the platform's translational and rotational motion.

[0025] Mode Switching: Real-time monitoring of the lithography machine's operating status signals. When the equipment is in standby or during wafer change intervals, the system automatically switches to energy-saving and cleaning mode. The active vibration damping module reduces the control bandwidth, the fan filter unit operates at reduced speed, and the directional adjustable airflow arm resets to standby. Before the equipment resumes operation, it smoothly switches back to high-performance mode. To achieve mode switching, the system has a built-in intelligent mode switching mechanism based on equipment status perception and disturbance prediction. The central coordinating controller obtains the operating status of the precision process equipment in real time through the SECS / GEM communication protocol. At the same time, it predicts future commands by combining the equipment disturbance characteristic model and uses real-time feedback from the sensor network as the basis for safety verification. When it is determined that the equipment has entered a low-dynamic stage such as standby or wafer change intervals and there is no high-intensity disturbance prediction for 500ms, the system automatically switches to energy-saving and cleaning mode. At this time, the active vibration damping module uses a linear load reduction curve to reduce the control bandwidth within 50-100ms. The bandwidth is reduced from 50-200Hz in high-performance mode to 5-10Hz, and the voice coil motor enters standby current state; the fan filter unit decelerates at a rate of 5-10% / s to the minimum speed required to maintain positive pressure, approximately 30-40% of the rated air volume; the adjustable directional airflow arm resets to the standby position without collision along the planned path, the composite nozzle shuts off the air supply, and the exhaust is reduced to the minimum flow rate to maintain positive pressure in the pipeline. When the equipment resumes operation 50ms before the system detects a high-intensity disturbance prediction command in advance, the system initiates a smooth pre-switch to high-performance mode. At this time, the active vibration damping module uses an S-shaped acceleration curve to restore the control bandwidth and gain to full value, the fan filter unit accelerates at a constant speed to the rated air volume, and the airflow arm moves to the predicted heat generation position in advance to standby. All switching processes are monitored in a closed loop by indicators such as residual vibration and particle concentration to ensure that full performance can be restored within 20ms without any impact, overshoot, or triggering any environmental alarms. This mechanism enables the system to reduce average energy consumption by more than 40% and extend actuator life by 2-3 times while ensuring an optimal process environment, and the switching process has zero interference with the process.

[0026] The central coordinating controller also superimposes the first feedforward control command, the second feedforward control command, and the feedback control command in the time domain to form a comprehensive control command that is finally output to each actuator. The first and second feedforward control commands are used to establish the reference values ​​for the system response. The output force reference of the active vibration reduction module and the airflow parameter reference of the microenvironment module are both confirmed by the feedforward command. The feedback control command is used to compensate for the residual error of the feedforward control and external sudden disturbances in real time, and to make dynamic fine adjustments based on the reference value state, so as to enable the system to achieve a balance between rapid response and precise stability.

[0027] The equipment disturbance model is the core data support for the central coordinating controller. It is a set of mathematical models describing the mapping relationship between the motion commands of precision manufacturing equipment and the multi-physics disturbances they induce. This model is pre-stored in the non-volatile memory of the central coordinating controller and continuously optimized and updated through machine learning algorithms during system operation. The input data of the model is the real-time motion commands of the equipment and their associated parameters, specifically including motion identifier, motion type, motion parameters, load parameters, and timing parameters. The load parameters are derived from the equipment parameter database, which contains parameters from the equipment's control system. The specific parameters of the motion identifier include the equipment ID and axis ID, with the data format being integer encoding. The motion type parameters include stepping, scanning, acceleration / deceleration, plate changing, and emergency stop, with the data format being enumerated values. The motion parameters include target position, velocity, acceleration, and jerk, with the data format being floating-point arrays. The specific parameters of the load parameters are the current moving parts. The mass and inertia are measured in floating-point format. The specific parameters of the timing parameters are the command trigger time and the expected duration, and the data format is timestamp. The output of the equipment disturbance model is multi-physics disturbance information, which is the direct basis for the system to perform feedforward control. The categories of output data include predicted vibration spectrum, predicted heat generation location, predicted pollution risk, and predicted timing window. The parameters of the predicted vibration frequency are the frequency domain amplitude and phase of each axis, and the data format is complex array, which is used to generate the first feedforward command. The specific parameters of the predicted heat generation location are the three-dimensional coordinates of the heat source and the power curve, and the data format is coordinate + time-varying function, which is used to generate the second feedforward command. The specific parameters of the predicted pollution risk are the particulate emission probability and the predicted particle size distribution, and the data format is probability value + distribution array, which is used to assist airflow control. The parameters of the predicted timing window are the disturbance start time, peak time, and end time, and the data format is time point, which is used for control timing coordination.

[0028] The equipment disturbance characteristic model adopts a hierarchical modular structure, consisting of three sub-models. The first is the vibration characteristic sub-model, which maps motion commands to a six-degree-of-freedom vibration spectrum at the equipment's load-bearing point. This sub-model includes a rigid body kinematics model, a structural transfer function matrix, and a modal superposition module. The rigid body kinematics module calculates theoretical inertial forces and moments based on the mass, inertia, and motion parameters of the moving parts of the equipment. The structural transfer function matrix describes the vibration transmission path from disturbance sources such as motors and guide rails to the equipment's load-bearing platform mounting point. The modal superposition module calculates the participation factors of each mode based on the equipment's finite element modal analysis structure. The output format of this sub-model is... This refers to the frequency domain transfer function or time domain impact response sequence. The finite element modal analysis structure refers to the dynamic numerical model of the equipment and its mounting platform established using the finite element method. The construction process includes discretizing the equipment's geometric model into millions of tiny elements, assigning material properties to each element such as density, elastic modulus, Poisson's ratio, and damping coefficient, defining the connection relationships between components such as bolted connections, guide rails and sliders, and flexible hinges, and obtaining the modal parameters of the equipment at each stage by solving the generalized eigenvalue problem. This includes modal frequencies indicating where the equipment is prone to resonance, modal shapes indicating the deformation of the equipment at each frequency, and descriptions of... The modal mass of each order of modes participates in the weighting. According to the modal superposition principle, the response of the equipment in actual motion can be expressed as a linear combination of the modal arrays of each order. The second sub-model is the thermal characteristic sub-model, which is used to map motion commands to the heat generation location and power curve of the internal heat source of the equipment. This sub-model includes a motor thermal model, a frictional thermal model, and a heat diffusion model. The motor thermal model calculates the heat generation of copper and iron losses based on motor current, speed, and torque. The frictional thermal model calculates the frictional heat generation based on the guide rail friction system and motion speed. The heat diffusion model predicts the heat diffusion range and time constant of the heat source location based on the heat conduction path of the equipment structure. The output shape of this sub-model is... The formula is a function of the heat source coordinates and power changing with time. The third sub-model is the pollution characteristic sub-model. The function of this model is to map motion commands into the probability and intensity of particulate emission. The structure of this sub-model includes a friction dust generation model, an airflow disturbance model, and a historical statistical model. The friction dust generation model calculates the particulate generation rate based on the material contact pressure and relative velocity of the relatively moving parts. The airflow disturbance model calculates the probability of particulates being entrained based on the velocity of the moving parts and the surrounding airflow field. The historical statistical model establishes a statistical regression model based on historical ion counter data of the same type of motion. The output of this sub-model is the particulate emission probability and the predicted particle size distribution.

[0029] The equipment disturbance characteristic model is obtained through a combination of four methods. The first method is theoretical modeling and simulation training. During the system design phase, based on the equipment's CAD model and dynamic parameters, virtual calibration is performed using multibody dynamics simulation software. Step one is to establish a rigid-flexible coupled dynamic model of the equipment; step two is to input a standard motion command set, such as a 10mm step and a 50mm / s scan; step three is to simulate the vibration response and motor heating power; and step four is to store the "command-response" data in a database as the initial version of the model. The second method is experimental calibration training. During the system installation and commissioning phase, the model is calibrated through controlled experiments. Step one is to deploy high-precision sensors on the equipment's platform and key components; step two is to have the control system execute standard motion commands one by one; step three is to synchronously collect vibration response and temperature change data; step four is to compare the measured data with the simulated prediction data and calculate the error; and step five is to use system discrimination algorithms such as subspace identification and neural networks to correct the model parameters and minimize the prediction error. The third method is online self-learning training. During system operation, the model is continuously optimized through machine learning algorithms. The first step is to record the entire process data for each execution of motion commands, including command parameters, feedforward commands, and sensor measured responses. The second step is to use the "command-measured response" as new training samples. The third step is to periodically call the offline training program and update the sub-models using vibration, thermal, and pollution model algorithms. The vibration model uses recursive least squares to update the transfer function matrix, the thermal model uses Gaussian process regression to update the heat generation location prediction, and the pollution model uses Bayesian to update the particulate emission probability. The fourth step is to verify the prediction accuracy of the updated model. If there is an improvement, the original model is replaced. The fourth method is transfer learning training. For multiple devices of the same model, transfer learning can be used to quickly build a model. The first step is to complete the full experimental calibration on one device and establish the source domain model. The second step is to conduct only a small number of rapid tests on other devices of the same model. The third step is to use the transfer learning algorithm to adapt the source domain model to the target device. The fourth step is to gradually replace it with a proprietary model as the target device's operating data accumulates.

[0030] The complete collaborative control process of the system is as follows: 1. In the disturbance prediction stage, the main control system of the lithography machine sends a signal to the workpiece stage at a speed of 2 m / s. 2 The motion command, "acceleration step 10mm," is synchronously captured by the central coordinating controller. By querying the internal disturbance characteristic model, the controller obtains the following information: the motion will generate a periodic vibration force with a main frequency of 30Hz and an amplitude of 20nm in the X direction; the linear motor coil of the worktable will generate an instantaneous heat load of 15W at coordinates x=150, y=200, z=50 after 3ms; the disturbance of the moving parts may generate a small amount of 0.1μm particles. 2. Feedforward Coordinated Execution Phase: This phase includes vibration control and airflow control. In the vibration control phase, the controller sends a pre-calculated reverse force current waveform with the same amplitude and opposite phase as the predicted force spectrum to the four active vibration isolation units. Each voice coil motor immediately generates a precise Lorentz force to push the equipment's load-bearing platform to counteract the vibration and bring the platform to a standstill. In the airflow control phase, the controller moves the adjustable directional airflow arm to the (150, 200, 50) coordinate point along the fastest path, aligns the composite nozzle with the heat source, sets the supply air velocity to 0.8 m / s, and the exhaust air volume to 15 L / min. At the same time, the rotation speed of the fan filter unit directly above this area is increased by 5%. 3. Feedback Fine-tuning Stage: This stage includes accelerometer feedback, thermal imaging feedback, and particle counter feedback. In the accelerometer feedback stage, the accelerometer on the equipment platform detects a residual rotational vibration of 0.5 nm. The feedback controller immediately calculates the compensation torque and adjusts the output distribution of the four active vibration isolation units, suppressing the residual vibration to below 0.1 nm within 0.2 ms. In the thermal imaging feedback stage, the thermal imager detects that the actual temperature rise is slightly higher than the model prediction, and the hot plume has an upward diffusion trend. At this time, the controller immediately dynamically increases the airflow velocity of the composite nozzle to 1.0 m / s and fine-tunes the nozzle pitch angle. Based on this, the hot spot will be completely suppressed within 1.5 s. For particle counter feedback, the real-time reading of the particle counter is always 0, and the system maintains the current capture parameters unchanged.

[0031] 4. All data from the entire collaborative process, including motion commands, model predictions, actual commands from each actuator, and sensor response curves, are stored in the historical database. After production ends on the same day, the central collaborative controller calls machine learning algorithms to train the data offline and automatically fine-tunes some gain coefficients in the disturbance feature model, thereby further improving the prediction accuracy of the next similar action.

[0032] For the system described above, the specific methods include the following steps: S1. Through the communication interface between the central coordinating controller and the equipment's control system, the next motion command that the equipment is about to execute is obtained in real time. This command includes parameter information such as motion type, motion axis, acceleration, velocity and displacement. S2. The central coordinating controller calls the internally stored equipment disturbance characteristic model and, based on the motion command obtained in step S1, predicts the multiphysics disturbance information that the command will trigger. The multiphysics disturbance information includes at least the following: Predicting the vibration spectrum: describing the amplitude and phase distribution of vibration in the frequency domain; Predicted heat generation location information: describes the three-dimensional spatial coordinates of the heat load source and the power variation curve over time; S3. Execute the first feedforward control. The central coordinating controller generates the corresponding reverse force spectrum based on the predicted vibration spectrum, converts it into a time-domain current waveform command, and sends it to the voice coil motor servo driver of the active vibration reduction module. Each voice coil motor generates a precise Lorentz force, which drives the equipment bearing platform to actively generate a movement that is opposite to the predicted vibration direction and equal in amplitude, thereby achieving feedforward cancellation of vibration. S4. Based on the predicted heat generation location information, the central coordinating controller generates a second feedforward control command and sends it to the microenvironment management module, which also includes: S4.1 Control the adjustable directional airflow arm to move to the three-dimensional spatial coordinates corresponding to the predicted heat generation location information; S4.2 Adjust the air supply speed and exhaust volume of the composite nozzle at the end of the adjustable directional airflow arm. At the same time, it is necessary to adjust the fan speed of the fan filter unit above the corresponding area to form a directional airflow barrier and negative pressure capture zone around the heat source, suppress the diffusion of the heat plume and capture particulate pollutants that may be emitted. S5. The central coordinating controller acquires real-time monitoring data from the sensor network at a high sampling rate, including vibration data from the accelerometer of the active vibration isolation unit and the non-contact vibration probe, temperature field data from the thermal imager, and pollutant concentration data from the particle counter. S6. The central coordinating controller compares the real-time monitoring data obtained in step S5 with the preset expected target values, such as zero vibration, constant temperature, and zero particles, and calculates the deviation. Based on the deviation, a feedback control command is generated and sent to the active vibration reduction module and the micro-environment management module. For the active vibration reduction module, the output force of each voice coil motor is finely adjusted to compensate for the residual error of the feedforward control and external sudden disturbances. For the micro-environment management module, the nozzle angle, air supply speed, exhaust air volume, and fan speed of the adjustable directional airflow arm are finely adjusted to accurately maintain the thermal and flow field stability of the core working area. S7 The central coordinating controller records the data of the entire process of each motion command execution to the historical database. The data must include at least motion command parameters, model prediction values, feedforward and feedback control commands, measured response curves of each sensor, and final environmental stability performance indicators. The system will periodically call machine learning algorithms to train the historical data and automatically optimize and update the parameters of the equipment disturbance characteristic model, so as to continuously improve the subsequent prediction and control accuracy.

[0033] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. An active vibration reduction and intelligent airflow coordinated control system for cleanrooms, characterized in that: It includes an equipment support platform, an active vibration damping module, a micro-environment management module, a sensor network, and a central collaborative controller. The active vibration damping module is located below the equipment support platform and is used to counteract the vibrations transmitted to the equipment support platform. The microenvironment management module includes a core working area surrounding the equipment and a support frame decoupled from the vibration of the equipment's carrying platform. The microenvironment management module is used to provide directional clean airflow to the core working area and capture pollutants. The sensor network is used to detect the vibration of the equipment carrying platform and the environmental state parameters of the core working area. The environmental state parameters include at least temperature field parameters and pollutant concentration parameters, so as to obtain real-time monitoring data. The central coordinating controller is connected to the active vibration reduction module, the micro-environment management module, the sensor network, and the control system signals of the equipment. The central coordinating controller is configured as follows: Store and recall preset microenvironment expected targets, including vibration limits, temperature fluctuation limits, and particle concentration limits; Receive real-time motion instructions from the control system. The real-time motion instructions include motion types and motion parameters that describe the actions that the device is about to perform. The real-time motion instructions are automation instructions within the control system. Based on its internally stored equipment disturbance feature model, the system predicts the multi-physics disturbance information that will be triggered according to real-time motion commands. The multi-physics disturbance information includes at least the predicted vibration spectrum and the predicted heat generation location information. Based on the predicted vibration spectrum, a first feedforward control command is generated and sent to the active vibration reduction module, so that the active vibration reduction module generates a canceling force with the same amplitude and opposite phase as the predicted vibration spectrum. Based on the predicted heat generation location information, a second feedforward control command is generated and sent to the microenvironment management module to adjust the clean airflow parameters at the corresponding location. Receive real-time monitoring data from the sensor network, and calculate the deviation between the real-time monitoring data and the expected target; Feedback control commands are generated based on the deviation between the real-time monitoring data and the expected target, and simultaneously sent to the active vibration reduction module and the micro-environment management module to coordinately adjust the output force of the active vibration reduction module and the airflow parameters of the micro-environment management module until the deviation converges to a preset threshold range.

2. The active vibration reduction and intelligent airflow coordinated control system for cleanrooms according to claim 1, characterized in that: The active vibration reduction module includes several active vibration isolation units. The equipment support platform is set on top of several active vibration isolation units. The active vibration isolation unit includes an air spring, a voice coil motor actuator, and a high-precision accelerometer.

3. The active vibration reduction and intelligent airflow coordinated control system for cleanrooms according to claim 2, characterized in that: The stator of the voice coil motor actuator is fixedly connected to the top plate of the air spring, the mover of the voice coil motor constitutes the top output end of the active vibration isolation unit, and the high-precision accelerometer is fixedly connected to the mover of the voice coil motor actuator to measure the acceleration of the output end.

4. The active vibration reduction and intelligent airflow coordinated control system for cleanrooms according to claim 3, characterized in that: The support frame is provided with several fan filter units located above the core working area. The fan filter units are used to provide clean airflow vertically downward. At least one adjustable directional airflow arm is provided around the fan filter units. The end of the adjustable directional airflow arm is provided with a composite nozzle.

5. The active vibration reduction and intelligent airflow coordinated control system for cleanrooms according to claim 4, characterized in that: The composite nozzle is internally divided into an air supply duct and an exhaust duct. The air supply duct is used to deliver airflow to the predicted heat generation location to form an airflow barrier, and the exhaust duct is used to form a negative pressure trapping zone around the airflow barrier.

6. The active vibration reduction and intelligent airflow coordinated control system for cleanrooms according to claim 5, characterized in that: The sensor network includes a vibration sensor, a thermal imager, and a particle counter. The vibration sensor is configured as a high-precision accelerometer and a vibration probe. The thermal imager is fixedly connected to the support frame, and its field of view covers the core working area. The isodynamic sampling port of the particle counter is integrated into the composite nozzle of the adjustable directional airflow arm.

7. The active vibration reduction and intelligent airflow coordinated control system for cleanrooms according to claim 1, characterized in that: The central coordinating controller is further configured to superimpose the first feedforward control command, the second feedforward control command, and the feedback control command in the time domain to form a final output integrated control command. The first feedforward control command and the second feedforward control command are used to establish a reference value for the system response, and the feedback control command is used to compensate for the residual errors of the first feedforward control command and the second feedforward control command as well as external sudden disturbances.

8. A method for active vibration reduction and intelligent airflow coordinated control of a cleanroom for implementing the system according to any one of claims 1-7, characterized in that: The method includes: S1. Obtain real-time motion commands from the equipment control system; S2. Based on the equipment disturbance feature model pre-stored in the central coordinating controller, predict the multi-physics disturbance information that will be triggered according to real-time motion commands. The multi-physics disturbance information includes at least the predicted vibration spectrum and the predicted heat generation location information. S3. Generate and execute the first feedforward control command to the active vibration reduction module to actively generate a counteracting force based on the predicted vibration spectrum; S4. Generate and execute a second feedforward control command to the microenvironment management module to adjust the clean airflow parameters at the corresponding location based on the predicted heat generation location information; S5. Acquire real-time monitoring data from the sensor network; S6. Generate feedback control commands based on the deviation between real-time monitoring data and expected targets, so as to coordinate the output force of the active vibration reduction module and the airflow parameters of the microenvironment management module. Step S4 further includes: S4.1 Control the adjustable directional airflow arm to move to the coordinates corresponding to the predicted heat generation location information; S4.2 Adjust the supply air velocity and exhaust air volume of the composite nozzle at the end of the adjustable directional airflow arm to suppress the predicted hot plume and capture the emitted pollutants.

Images