A wraparound intelligent braking system for CNC rotary tables based on AI predictive control

Abstract

This invention provides an AI-predictive control-based intelligent braking system for a CNC rotary table, belonging to the fields of intelligent manufacturing and CNC technology. It solves the problems of lag response, lack of intelligent prediction and adaptive capabilities, insufficient control precision, and long-term performance degradation found in existing CNC rotary table braking systems. This AI-predictive control-based intelligent braking system for a CNC rotary table includes a state acquisition module, an AI prediction module, a pressure control module, a wraparound braking execution structure, and a feedback correction module. By introducing an AI prediction mechanism and a self-correcting feedback loop, this invention achieves intelligent, adaptive, and highly responsive braking control, improving the positioning accuracy, braking stability, and overall energy efficiency of the CNC rotary table. It boasts advantages such as significantly improved response speed, high positioning accuracy, strong adaptive capability, smooth braking process, and high level of intelligence.

Description

Technical Field

[0001] This invention belongs to the field of intelligent manufacturing and CNC technology, and relates to an intelligent braking system for a wraparound CNC turntable based on AI predictive control. Background Technology

[0002] In high-precision CNC equipment (such as five-axis machining centers), the CNC rotary table is a core functional component, and its braking performance (response speed, positioning accuracy, and stability) directly affects the machining quality. Currently, mainstream CNC rotary table braking systems mainly rely on mechanical clamping, hydraulic / pneumatic piston clamping, or magnetic powder braking. Among these, hydraulic / pneumatic driven brakes or chuck structures (mechanical braking structures with a similar "hugging" characteristic) are relatively common solutions, but they still have the following problems:

[0003] Response lag: Existing systems mostly rely on preset thresholds (such as position or speed thresholds) to trigger braking. When the turntable's operating state (such as speed or load) changes suddenly, it cannot predict and initiate braking actions in time, resulting in delayed braking response, long braking distance, and affecting processing accuracy and equipment safety.

[0004] Lack of intelligent prediction and adaptive capabilities: Traditional control is based on real-time status and lacks effective prediction of the future movement trend of the turntable (such as overshoot and load variation). It is difficult to achieve early intervention of braking action and dynamic optimization adjustment of clamping force. The control parameters are fixed and cannot adapt to different working conditions (such as high speed, low speed, heavy load, and light load).

[0005] Insufficient control precision: When rotating at high speed or experiencing sudden load changes, existing systems have difficulty accurately controlling the braking process, which can easily lead to braking overshoot or undershoot, resulting in increased positioning errors.

[0006] Long-term performance degradation: Existing systems generally lack self-learning and self-correction mechanisms, which cannot effectively compensate for performance drift caused by wear of braking mechanisms (such as friction pads), changes in oil characteristics, and environmental factors (such as temperature). This leads to a decline in system braking performance over time, affecting the long-term operational stability of the equipment.

[0007] Therefore, designing an AI-based predictive control-based intelligent braking system for a wraparound CNC rotary table is of great significance. Summary of the Invention

[0008] The purpose of this invention is to address the aforementioned problems in the prior art by providing an AI predictive control-based intelligent braking system for a wraparound CNC turntable.

[0009] The objective of this invention can be achieved through the following technical solution: an AI predictive control-based intelligent braking system for a wraparound CNC turntable, characterized in that it comprises:

[0010] The status acquisition module is used to acquire the operating parameters of the CNC rotary table in real time, collect the real-time status of the system, and provide a data basis for prediction. The operating parameters include angle, speed, load current and temperature.

[0011] The AI ​​prediction module, connected to the status acquisition module, is used to predict the short-term motion trend and braking demand of the CNC rotary table based on the acquired historical and current operating parameters using a machine learning model, and output the predicted braking start timing and clamping force control parameters.

[0012] The pressure control module, connected to the AI ​​prediction module, is used to generate a control signal based on the predicted braking start timing and clamping force control parameters.

[0013] The ring-shaped brake actuator structure is connected to the pressure control module and is controlled by the control signal. The ring-shaped brake actuator structure includes an elastic thin-walled brake ring 2 and a pressure oil chamber. By applying controlled oil pressure to the pressure oil chamber, the elastic thin-walled brake ring 2 is elastically deformed to tighten or loosen the turntable rotating shaft.

[0014] The feedback correction module, connected to the state acquisition module and the AI ​​prediction module, is used to collect actual braking result data, compare it with the expected prediction result of the AI ​​prediction module and calculate the error, and adjust the parameters of the machine learning model online based on the error to achieve self-learning and adaptive optimization of the system.

[0015] In the aforementioned AI-based predictive control-based intelligent braking system for a wraparound CNC turntable, the machine learning model is a Long Short-Term Memory (LSTM) network, an XGBoost algorithm, or a reinforcement learning model. Using the machine learning model, based on real-time collected turntable operating status data, the system actively predicts the braking demand and optimizes the braking control parameters for the short term.

[0016] In the aforementioned AI predictive control-based intelligent braking system for a wraparound CNC rotary table, the pressure control module includes an electromagnetic proportional valve for precisely adjusting the magnitude and rate of change of the oil pressure entering the pressure oil chamber of the wraparound brake actuator. By precisely controlling the flow rate and pressure of the pressure oil entering the brake ring oil chamber, stepless, rapid, and precise adjustment of the clamping force is achieved.

[0017] In the aforementioned AI-predictive control-based intelligent braking system for a wraparound CNC rotary table, the pressure control module is configured to control the wraparound braking actuator to execute sequentially based on the predicted clamping force control parameters:

[0018] Pre-clamping stage: Quickly apply initial clamping force to make slight contact between the elastic thin-walled brake ring and the turntable rotation axis, quickly establish initial frictional contact, eliminate gaps and begin deceleration;

[0019] Full clamping stage: According to the predicted optimization target, control the oil pressure to rise to the target value or dynamically adjust it as needed, accurately control the oil pressure rise curve, so that the clamping force rises smoothly and quickly to the target value, and generate a large enough braking torque to make the turntable stop completely within the target time.

[0020] In the aforementioned AI predictive control-based intelligent braking system for a wraparound CNC rotary table, the feedback correction module adjusts the weight parameters of the machine learning model online using a gradient descent algorithm, optimizes the parameters of the AI ​​predictive model, and achieves self-learning and long-term performance stability of the system.

[0021] A method for intelligent braking of a wraparound CNC rotary table based on AI predictive control, characterized by the following steps:

[0022] S1: Utilize machine learning models to collect real-time operating parameters of the CNC rotary table, including angle, speed, load current, and temperature.

[0023] S2: Based on the collected historical and current operating parameters, the machine learning model is used to predict the short-term motion trend and braking demand of the CNC rotary table in the future, and the predicted braking start time and clamping force control parameters are obtained.

[0024] S3: Based on the predicted braking start timing and clamping force control parameters, a control signal is generated and the ring-type brake actuator is driven to perform braking action, so as to control the action timing and clamping force application process of the ring-type brake actuator, realize advanced intervention and process optimization, wherein the braking action includes using controlled oil pressure to cause the elastic thin-walled brake ring of the ring-type brake actuator to produce elastic deformation to clamp the turntable rotation shaft.

[0025] S4: Collect actual braking result data;

[0026] S5: Compare the actual braking results with the predicted expected results and calculate the error;

[0027] S6: Based on the error, adjust the parameters of the machine learning model online to achieve self-learning and adaptive updating of the model, thereby improving the accuracy of subsequent predictions and the long-term stability of braking control performance.

[0028] Compared with existing technologies, this AI predictive control-based intelligent braking system for a wraparound CNC rotary table has the following significant advantages:

[0029] 1. Significantly improved response speed: AI prediction enables proactive initiation and optimization of braking actions, significantly shortening the response time from identifying the need to effective braking, especially under conditions such as sudden changes in speed or load.

[0030] 2. High positioning accuracy: By combining predictive control with hydraulic dynamic adjustment, precise control of the braking process is achieved, which can more accurately control the braking process, effectively reduce overshoot and under-braking, and significantly improve the final positioning accuracy (for example, the error between the target position and the actual stopping position is significantly reduced).

[0031] 3. Strong adaptability: The system can automatically adjust control parameters according to the prediction results to adapt to different speeds and load conditions. The online feedback correction mechanism can continuously learn and compensate for system characteristic drift such as friction plate wear and temperature changes, ensuring long-term stability and accuracy.

[0032] 4. Smooth braking process: The phased clamping force control strategy (such as pre-clamping and full clamping) is adopted to dynamically adjust the clamping force, effectively reduce braking impact, improve the smoothness of the braking process, protect the mechanical structure and improve the processing quality.

[0033] 5. High level of intelligence: The system integrates real-time status perception, motion trend prediction, control decision optimization and online self-learning capabilities, realizing intelligent braking control and significantly improving the overall level of intelligence. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating the intelligent braking system for a wraparound CNC rotary table based on AI predictive control.

[0035] Figure 2 This is a schematic diagram of the turntable rotation axis of the wraparound CNC turntable intelligent braking system based on AI predictive control.

[0036] Figure 3 This is a schematic diagram of the pressure oil chamber of an AI-based predictive control-based intelligent braking system for a wraparound CNC rotary table.

[0037] Figure 4 This is a flowchart of the AI ​​control strategy for an AI-based predictive control-based intelligent braking system for a wraparound CNC rotary table.

[0038] In the diagram, 1. Turntable rotation axis; 101. Status acquisition module; 102. AI prediction module; 103. Pressure control module; 104. Ring-type brake actuation structure; 105. Feedback correction module; 2. Elastic thin-walled brake ring; 201. Pressure oil chamber; 3. Brake ring shell; 301. Channel hole one; 302. Channel hole two; 303. Channel hole three; 304. Channel hole four; 4. Brake base; 401. Oil inlet; 402. Oil outlet; 5. O-ring one; 6. O-ring two. Detailed Implementation

[0039] The following are specific embodiments of the present invention, which are described in conjunction with the accompanying drawings. However, the present invention is not limited to these embodiments.

[0040] like Figure 1-4 As shown, this AI predictive control-based wraparound CNC turntable intelligent braking system includes:

[0041] Status acquisition module 101: It consists of sensors installed on the turntable (such as photoelectric encoder to measure angular displacement / velocity, current transformer to measure motor load current, and temperature sensor), and is used to collect the turntable's angular displacement, speed, load current, system temperature and other pipe operating parameters in real time.

[0042] AI prediction module 102: A software module running in the control system (such as PLC, industrial computer or embedded processor), which adopts a trained machine learning model (preferably LSTM time series prediction model, but can also be XGBoost, Transformer or other applicable time series models or reinforcement learning agents). Its input is the historical and real-time data stream provided by the state acquisition module 101, and its output is the predicted braking demand (e.g.: the braking needs to be started within XX milliseconds in the future, the predicted stopping position, the required initial clamping force, and the clamping force change curve suggestion).

[0043] Pressure control module 103: Receives instructions from AI prediction module 102 and converts them into precise control signals for the brake actuator. A preferred embodiment is a hydraulic / pneumatic pressure regulating system composed of an electromagnetic proportional valve. By precisely controlling the flow rate and pressure of the pressure oil entering the brake ring oil chamber, stepless, rapid, and precise adjustment of the clamping force can be achieved. An alternative embodiment may include a mechanical actuator driven by a servo motor. Predictive control combined with dynamic pressure regulation can more accurately control the braking process, effectively reduce overshoot and under-braking, and significantly improve the final positioning accuracy. For example, the error between the target position and the actual stopping position is significantly reduced.

[0044] The ring-type brake actuator structure 104 includes an elastic thin-walled brake ring 2 and a pressure oil chamber 201. By applying controlled oil pressure to the pressure oil chamber 201, the elastic thin-walled brake ring 2 is elastically deformed to tighten or loosen the turntable rotating shaft 1.

[0045] Feedback correction module 105: After the braking action is completed, it collects the actual braking results, such as the actual stopping position read by the encoder and the time-speed curve of the braking process.

[0046] The actual braking results are compared with the previous predictions of the AI ​​prediction module 102 to calculate the error (such as position error and velocity curve fitting error). Using this error information, the model parameters in the AI ​​prediction module 102 are fine-tuned online through algorithms such as backpropagation, continuously optimizing the prediction accuracy of the model. The feedback correction mechanism can continuously learn and compensate for system changes (such as wear and temperature effects) to ensure the stability and accuracy of long-term operation.

[0047] The ring-shaped brake actuator structure 104 consists of:

[0048] Turntable rotation axis 1: The object being braked;

[0049] Elastic thin-walled brake ring 2: Core component, typically made of a thin-walled ring structure from a highly elastic metal material (such as spring steel). See details. Figure 3 ;

[0050] Brake ring housing 3: Fixing component;

[0051] Brake base 4: Fixing component.

[0052] The upper end of the elastic thin-walled brake ring 2 is connected to the brake ring housing 3 by screws, and the lower end of the elastic thin-walled brake ring 2 is connected to the brake base 4. The brake base 4 is connected to the brake ring housing 3 by screws, so that the elastic thin-walled brake ring 2, the brake ring housing 3 and the brake base 4 form a rigid whole.

[0053] The upper and lower ends of the elastic thin-walled brake ring 2 have sealing grooves on their outer walls, and the middle section of the elastic thin-walled brake ring 2 has an annular groove on its inner wall. This groove forms a pressure oil chamber 201. O-ring 1 5 and O-ring 2 6 are respectively installed in the upper and lower sealing grooves to ensure the sealing of the pressure oil chamber 201.

[0054] Oil passage: Pressurized oil passes through inlet 401, channel hole 1 301, and channel hole 2 302 in sequence to reach pressure oil chamber 201; it can also pass through channel hole 3 303 and channel hole 4 304 in sequence to reach outlet 402 and flow back to the oil tank. Channel hole 1 301, channel hole 2 302, channel hole 3 303 and channel hole 4 304 are all inside brake ring housing 3.

[0055] Working principle:

[0056] Normal operation: The outer circle of the turntable rotation shaft 1 and the inner wall of the elastic thin-walled brake ring 2 maintain a design gap (e.g., 0.1mm) to ensure the turntable rotates freely.

[0057] Braking process: When braking is required, pressurized oil enters the pressure oil chamber 201 through the above-mentioned oil passage. The oil pressure acts on the thin-walled part of the elastic thin-walled brake ring 2 in the pressure oil chamber 201, forcing the thin-walled area to produce inward elastic deformation, so that the inner wall of the elastic thin-walled brake ring 2 tightly hugs the outer circle of the turntable rotation shaft 1, generating a strong frictional braking torque, which makes the turntable stop quickly.

[0058] Release the brake: When the turntable needs to resume rotation, the oil pressure in the pressure oil chamber 201 is unloaded (the pressure oil flows back to the oil tank through the channel hole 303 and the channel hole 404, and reaches the oil outlet 402). The elastic thin-walled brake ring 2 returns to its original shape by relying on the elastic restoring force of its own material. The gap (0.1mm) between the inner wall and the outer circle of the turntable rotation shaft 1 reappears, the friction disappears, and the turntable can rotate freely.

[0059] AI control strategy process, such as Figure 4 :

[0060] S1: Status acquisition module 101: Real-time continuous acquisition of data such as angular displacement θ, rotational speed ω, load current I, and temperature T.

[0061] S2: AI prediction module 102: Inputs the collected real-time data stream (including historical segments) into the trained LSTM model;

[0062] The model outputs prediction results, such as: predicting whether an event requiring emergency braking will occur within a short future window (e.g., Δt=300ms) (e.g., predicting tool breakage based on load current mutation rate dI / dt).

[0063] Predict the future trajectory of the turntable (speed changes, expected stopping position);

[0064] Predict the required braking initiation time, initial clamping force (e.g., percentage of maximum clamping force, 30% F_max), and the time or curve for the clamping force to reach the target value.

[0065] If braking is predicted, a request to pause machining is immediately sent to the CNC system, and a braking command containing detailed control parameters (such as target pressure and pressure rise slope) is sent to the pressure control module 103.

[0066] S3: Pressure control module 103 (pressure control execution):

[0067] Pre-clamping stage (rapid response): The pressure control module 103 (proportional valve system) opens rapidly according to the command, allowing pressure oil to quickly enter the pressure oil chamber 201 of the elastic thin-walled brake ring 2, pushing the inner wall of the brake ring to make slight contact with the rotating shaft 1 of the turntable (applying pre-clamping force, such as 30% F_max). This stage is designed to quickly establish initial frictional contact, eliminate gaps and begin deceleration.

[0068] Full clamping stage (precise control): After or simultaneously with the pre-clamping, the pressure control module 103 precisely controls the oil pressure rise curve according to the optimization target predicted by AI (such as avoiding overshoot), so that the clamping force rises smoothly and quickly to the target value (100% F_max or a value dynamically adjusted according to the prediction), generating a sufficiently large braking torque to make the turntable stop completely within the target time (e.g., ≤80ms). This stage can avoid the impact caused by rough braking.

[0069] S4: Feedback correction module 105: After braking is completed, the encoder accurately records the actual stop position of the turntable (θ_stop_actual) and the speed change curve (ω(t)) during the entire braking process.

[0070] The actual stopping position θ_stop_actual is compared with the target stopping position θ_stop_predicted predicted by the AI ​​prediction module 102, and the position error Δθ is calculated.

[0071] Compare the actual speed change curve ω(t)_actual with the expected speed curve ω(t)_predicted predicted by the AI ​​prediction module 102 (e.g., calculate the mean square error).

[0072] Based on the above error information (Δθ, velocity curve error), the feedback correction module 105 uses an online learning algorithm (e.g., parameter update based on gradient descent) to fine-tune the LSTM model parameters in the AI ​​prediction module 102. Through this continuous "prediction-execution-feedback-learning" closed loop, the prediction accuracy of the model is continuously improved over time, and the system's adaptive capability is enhanced.

[0073] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

[0074] Although this document uses a number of technical terms, the possibility of using other terms is not excluded. These terms are used merely for the convenience of describing and explaining the essence of the invention; interpreting them as any additional limitation would be contrary to the spirit of the invention.

Claims

1. A wraparound intelligent braking system for a CNC rotary table based on AI predictive control, characterized in that, include: The status acquisition module (101) is used to acquire the operating parameters of the CNC rotary table in real time, collect the real-time status of the system, and provide a data basis for prediction. The operating parameters include angle, speed, load current and temperature. The AI ​​prediction module (102) is connected to the status acquisition module (101) and is used to predict the short-term motion trend and braking demand of the CNC rotary table based on the acquired historical and current operating parameters using a machine learning model, and output the predicted braking start time and clamping force control parameters. The pressure control module (103) is connected to the AI ​​prediction module (102) and is used to generate a control signal based on the predicted braking start timing and clamping force control parameters. The ring-type brake actuator (104) is connected to the pressure control module (103) and is controlled by the control signal. The ring-type brake actuator (104) includes an elastic thin-walled brake ring (2) and a pressure oil chamber (201). By applying controlled oil pressure to the pressure oil chamber (201), the elastic thin-walled brake ring (2) is elastically deformed to tighten or loosen the turntable rotating shaft (1). The feedback correction module (105) is connected to the state acquisition module (101) and the AI ​​prediction module (102). It is used to collect actual braking result data, compare it with the expected prediction result of the AI ​​prediction module (102) and calculate the error, and adjust the parameters of the machine learning model online based on the error to realize the self-learning and adaptive optimization of the system.

2. The intelligent braking system for a wraparound CNC rotary table based on AI predictive control according to claim 1, characterized in that, The machine learning model is a Long Short-Term Memory (LSTM) network, XGBoost algorithm, or reinforcement learning model. Based on real-time collected turntable operating status data, the machine learning model actively predicts the braking demand and optimizes the braking control parameters in the short term.

3. The intelligent braking system for a wraparound CNC rotary table based on AI predictive control according to claim 1, characterized in that, The pressure control module (103) includes an electromagnetic proportional valve, which is used to precisely adjust the magnitude and rate of change of the oil pressure entering the pressure oil chamber (201) of the ring brake actuator structure (104). By precisely controlling the flow rate and pressure of the pressure oil entering the brake ring oil chamber, stepless, rapid and precise adjustment of the clamping force can be achieved.

4. The intelligent braking system for a wraparound CNC rotary table based on AI predictive control according to claim 1, characterized in that, The pressure control module (103) is configured to control the ring-type brake actuator (104) to perform the following sequentially based on the predicted clamping force control parameters: Pre-clamping stage: Quickly apply initial clamping force to make the elastic thin-walled brake ring (2) make slight contact with the turntable rotating shaft (1), quickly establish initial frictional contact, eliminate gaps and start deceleration; Full clamping stage: According to the predicted optimization target, control the oil pressure to rise to the target value or dynamically adjust it as needed, accurately control the oil pressure rise curve, so that the clamping force rises smoothly and quickly to the target value, and generate a large enough braking torque to make the turntable stop completely within the target time.

5. The intelligent braking system for a wraparound CNC rotary table based on AI predictive control according to claim 1, characterized in that, The feedback correction module (105) adjusts the weight parameters of the machine learning model online through the gradient descent algorithm, optimizes the parameters of the AI ​​prediction model, and realizes the system's self-learning and long-term performance stability.

6. A method for intelligent braking of a wraparound CNC rotary table based on AI predictive control according to any one of claims 1 to 5, characterized in that, Includes the following steps: S1: Utilize machine learning models to collect real-time operating parameters of the CNC rotary table, including angle, speed, load current, and temperature. S2: Based on the collected historical and current operating parameters, the machine learning model is used to predict the short-term motion trend and braking demand of the CNC rotary table in the future, and the predicted braking start time and clamping force control parameters are obtained. S3: Based on the predicted braking start timing and clamping force control parameters, a control signal is generated and the ring-type brake actuator (104) is driven to perform braking action, so as to control the action timing and clamping force application process of the ring-type brake actuator (104) to achieve advanced intervention and process optimization. The braking action includes using controlled oil pressure to cause elastic deformation of the elastic thin-walled brake ring (2) of the ring-type brake actuator (104) to clamp the turntable rotating shaft (1). S4: Collect actual braking result data; S5: Compare the actual braking results with the predicted expected results and calculate the error; S6: Based on the error, adjust the parameters of the machine learning model online to achieve self-learning and adaptive updating of the model, thereby improving the accuracy of subsequent predictions and the long-term stability of braking control performance.

Images