Photoetching machine motor system with self-adaptive control function

Abstract

The invention provides a photoetching machine motor system with a self-adaptive control function, which comprises a control module and a sensor, and is characterized in that the sensor acquires operation data of a motor in real time and transmits acquired operation parameters of the motor to the control module; the control module receives motor operation parameters collected by the sensor, and can automatically optimize control parameters through a self-adaptive control algorithm according to historical data and real-time feedback. Through a self-adaptive control algorithm, control parameters can be adjusted in real time, and the operation precision of the photoetching machine motor is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photolithography motors, and in particular to a photolithography machine motor system with an adaptive control function. Background Art

[0002] In the semiconductor manufacturing process, the photolithography machine is one of the key production equipment, and its performance directly affects the quality and output of chips. Traditional photolithography machine motor control systems mostly use fixed parameter control strategies, which are difficult to adapt to various changes in the production process, such as temperature fluctuations, load changes, etc., which limits the accuracy and efficiency of the photolithography machine. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a photolithography machine motor system with an adaptive control function, which is specifically achieved through the following technical solutions:

[0004] A photolithography machine motor system with adaptive control function comprises a control module and a sensor, wherein:

[0005] The sensor collects the operating data of the motor in real time and transmits the collected motor operating parameters to the control module;

[0006] The control module receives the motor operating parameters collected by the sensor, and can automatically optimize the control parameters according to historical data and real-time feedback through an adaptive control algorithm.

[0007] Optionally or preferably, the adaptive control algorithm includes the following:

[0008] Data collection: collect the working status and environmental condition data of the motor in real time through sensors;

[0009] Feature extraction: Use convolutional neural networks to build machine learning models and extract key features from collected data;

[0010] Model updating: Update the prediction model based on the extracted key features to more accurately describe the dynamic behavior of the system;

[0011] Control parameter optimization: Based on the updated model, the MPC algorithm is used to calculate the optimal control parameters;

[0012] Execution control: Apply the calculated control parameters to the motor system to adjust the motor's operating state.

[0013] Optionally or preferably, when a convolutional neural network is used for feature extraction, the following contents are included:

[0014] Data preprocessing: Clean the data collected by the sensor, remove noise and outliers; then standardize the data to ensure that different features have the same dimension, which is convenient for subsequent processing and makes it suitable as CNN input;

[0015] Convolution operation: use multiple convolution kernels to extract local features in the data;

[0016] Pooling operation: Reduce dimension through maximum pooling or average pooling while retaining important features;

[0017] Fully connected layer: All features are combined to form the final feature vector.

[0018] Optionally or preferably, when performing feature extraction, the cleaned data is divided into a training set, a validation set, and a test set in a ratio of 8:1:1.

[0019] Optionally or preferably, when the MPC algorithm is used to calculate the optimal control parameters, the following steps are repeated continuously to achieve continuous optimization control of the system:

[0020] S1. Establish a system model to describe the dynamic behavior of the lithography machine motor system:

[0021] x i+1 =Ax i +Bu i

[0022] Among them, x i represents the system state at the i-th moment; u i represents the control input at the i-th moment; A represents the system matrix; B represents the control input matrix;

[0023] S2. Based on the system model established in step S1, predict the state change of the photolithography motor system in the future. Specifically, assuming that the prediction step length is N, the state of the next N steps is calculated in an iterative manner:

[0024]

[0025]

[0026] This state can be expressed in matrix form as:

[0027]

[0028] Where U is the control input sequence for the next N-1 steps:

[0029]

[0030] S3. According to the above content and the set optimization goal, construct an objective function f(U):

[0031] f(U)=(XX ref ) T Q(XX ref )+U T RU

[0032] Among them, X ref is the reference state for the next N steps; Q and R are the weight matrices of state and control respectively;

[0033] Substituting the prediction model into the objective function, we get:

[0034] f(U)=(M+KU-X ref ) T Q(M+KU-X ref )+U T RU

[0035] Among them, M is the future state caused by the current state and external disturbance, and K is a matrix composed of the control input matrix;

[0036] S4. Solve the control input sequence U with the minimum objective function f(U). The control input sequence obtained is U * ;

[0037] S5. The control input sequence obtained by solving step S4 is U * The first element u i * As the control input at the current moment;

[0038] S6. At the next sampling moment, the prediction model is updated with the new measurement value, and the control input sequence with the minimum objective function is re-solved to obtain a new control sequence.

[0039] Based on the above technical solution, the following technical effects can be produced:

[0040] The present invention provides a photolithography machine motor system with an adaptive control function, and its beneficial effects include:

[0041] (1) High precision: Through the adaptive control algorithm, the control parameters can be adjusted in real time to improve the operating accuracy of the lithography machine motor;

[0042] (2) High efficiency: Optimized control parameters can reduce energy consumption and improve production efficiency;

[0043] (3) Strong adaptability: The system can adapt to different working conditions and environmental changes to ensure the stable operation of the lithography machine;

[0044] (4) Easy to maintain: The modular design of the system makes maintenance and upgrades more convenient. DETAILED DESCRIPTION

[0045] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0046] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Embodiment 1:

[0048] This embodiment provides a photolithography machine motor system with an adaptive control function, including a control module and a sensor, wherein:

[0049] The sensor collects the motor's operating data in real time and transmits the collected motor operating parameters to the control module; specifically, the operating data includes key parameters such as speed, position, and temperature. The data collection frequency must meet the real-time requirements of the control system;

[0050] The control module receives the motor operating parameters collected by the sensor, and can automatically optimize the control parameters according to historical data and real-time feedback through an adaptive control algorithm.

[0051] In this embodiment, the adaptive control algorithm includes the following contents:

[0052] Data collection: collect the working status and environmental condition data of the motor in real time through sensors;

[0053] Feature extraction: Use convolutional neural networks to build machine learning models and extract key features from collected data;

[0054] Model updating: Update the prediction model based on the extracted key features to more accurately describe the dynamic behavior of the system;

[0055] Control parameter optimization: Based on the updated model, the MPC algorithm is used to calculate the optimal control parameters;

[0056] Execution control: Apply the calculated control parameters to the motor system to adjust the motor's operating state.

[0057] In this embodiment, when a convolutional neural network is used for feature extraction, the following contents are included:

[0058] Data preprocessing: Clean the data collected by the sensor, remove noise and outliers; then standardize the data to ensure that different features have the same dimension, which is convenient for subsequent processing and makes it suitable as CNN input;

[0059] Convolution operation: use multiple convolution kernels to extract local features in the data;

[0060] Pooling operation: Reduce dimension through maximum pooling or average pooling while retaining important features;

[0061] Fully connected layer: All features are combined to form the final feature vector.

[0062] In this embodiment, when performing feature extraction, the cleaned data is divided into a training set, a validation set, and a test set in a ratio of 8:1:1.

[0063] In this embodiment, when the MPC algorithm is used to calculate the optimal control parameters, the following steps are repeated continuously to achieve continuous optimization control of the system:

[0064] S1. Establish a system model to describe the dynamic behavior of the lithography machine motor system:

[0065] x i+1 =Ax i +Bu i

[0066] Among them, x i represents the system state at the i-th moment; u i represents the control input at the i-th moment; A represents the system matrix; B represents the control input matrix;

[0067] S2. Based on the system model established in step S1, predict the state change of the photolithography motor system in the future. Specifically, assuming that the prediction step length is N, the state of the next N steps is calculated in an iterative manner:

[0068]

[0069] This state can be expressed in matrix form as:

[0070]

[0071] Where U is the control input sequence for the next N-1 steps:

[0072]

[0073] S3. According to the above content and the set optimization goal, construct an objective function f(U):

[0074] f(U)=(XX ref ) T Q(XX ref )+U T RU

[0075] Among them, X ref is the reference state for the next N steps; Q and R are the weight matrices of state and control respectively;

[0076] Substituting the prediction model into the objective function, we get:

[0077] f(U)=(M+KU-X ref ) T Q(M+KU-X ref )+U T RU

[0078] Among them, M is the future state caused by the current state and external disturbance, and K is a matrix composed of the control input matrix;

[0079] S4. Solve the control input sequence U with the minimum objective function f(U). The control input sequence obtained is U * ;

[0080] S5. The control input sequence obtained by solving step S4 is U * The first element u i * As the control input at the current moment;

[0081] S6. At the next sampling moment, the prediction model is updated with the new measurement value, and the control input sequence with the minimum objective function is re-solved to obtain a new control sequence.

[0082] The structures, functions and connection forms disclosed in this article can be implemented in other ways. For example, the embodiments described above are only illustrative, and there may be other installation methods, such as multiple components can be combined or integrated into another component; in addition, each functional component in each embodiment of this article can be integrated into one functional component, or each functional component can exist physically separately, or two or more functional components can be integrated into one functional component.

[0083] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims

1. A photolithography machine motor system with adaptive control function, characterized in that: It includes control module and sensor, among which: The sensor collects the operating data of the motor in real time and transmits the collected motor operating parameters to the control module; The control module receives the motor operating parameters collected by the sensor, and can automatically optimize the control parameters according to historical data and real-time feedback through an adaptive control algorithm.

2. The photolithography machine motor system with adaptive control function according to claim 1, characterized in that: The adaptive control algorithm includes the following: Data collection: collect the working status and environmental condition data of the motor in real time through sensors; Feature extraction: Use convolutional neural networks to build machine learning models and extract key features from collected data; Model updating: Update the prediction model based on the extracted key features to more accurately describe the dynamic behavior of the system; Control parameter optimization: Based on the updated model, the MPC algorithm is used to calculate the optimal control parameters; Execution control: Apply the calculated control parameters to the motor system to adjust the motor's operating state.

3. The photolithography machine motor system with adaptive control function according to claim 2, characterized in that: When using convolutional neural networks for feature extraction, the following are included: Data preprocessing: Clean the data collected by the sensor, remove noise and outliers; then standardize the data to ensure that different features have the same dimension, which is convenient for subsequent processing and makes it suitable as CNN input; Convolution operation: use multiple convolution kernels to extract local features in the data; Pooling operation: Reduce dimension through maximum pooling or average pooling while retaining important features; Fully connected layer: All features are combined to form the final feature vector.

4. The photolithography machine motor system with adaptive control function according to claim 3, characterized in that: When performing feature extraction, the cleaned data is divided into training set, validation set and test set in a ratio of 8:1:

1.

5. The photolithography machine motor system with adaptive control function according to claim 2, characterized in that: When using the MPC algorithm to calculate the optimal control parameters, the following steps are repeated continuously to achieve continuous optimization control of the system: S1. Establish a system model to describe the dynamic behavior of the lithography machine motor system: x i+1 =Ax i +Bu i Among them, x i represents the system state at the i-th moment; u i represents the control input at the i-th moment; A represents the system matrix; B represents the control input matrix; S2. Based on the system model established in step S1, predict the state change of the photolithography motor system in the future. Specifically, assuming that the prediction step length is N, the state of the next N steps is calculated in an iterative manner: This state can be expressed in matrix form as: Where U is the control input sequence for the next N-1 steps: S3. According to the above content and the set optimization goal, construct an objective function f(U): f(U)=(X-X ref ) T Q(X-X ref )+U T RU Among them, X ref is the reference state for the next N steps; Q and R are the weight matrices of state and control respectively; Substituting the prediction model into the objective function, we get: f(U)=(M+KU-X ref ) T Q(M+KU-X ref )+U T RU Among them, M is the future state caused by the current state and external disturbance, and K is a matrix composed of the control input matrix; S4. Solve the control input sequence U with the minimum objective function f(U). The control input sequence obtained is U * ; S5. The control input sequence obtained by solving step S4 is U * The first element u i * As the control input at the current moment; S6. At the next sampling moment, the prediction model is updated with the new measurement value, and the control input sequence with the minimum objective function is re-solved to obtain a new control sequence.

Images