A Method and System for Automatic Generation of Semiconductor Facility Layout Diagrams Based on Artificial Intelligence Models

Abstract

This invention discloses a method, system, equipment, and medium for automatically generating semiconductor plant layout diagrams based on an artificial intelligence model. The method includes receiving semiconductor plant layout diagram data; standardizing the received equipment data to convert it into a unified format; inputting the standardized data into an artificial intelligence model for vectorization processing to generate numerical vectors representing equipment characteristics; automatically calculating the optimal arrangement of machines within the semiconductor plant using a layout optimization algorithm based on the generated numerical vectors; automatically generating pipeline connection schemes between machines based on the machine arrangement results, including pipeline path planning and connection point location; and outputting the final generated semiconductor plant layout diagram. This invention significantly reduces the probability of errors in manual operation and improves the overall quality and efficiency of semiconductor plant layout design.

Description

Technical Field

[0001] This invention belongs to the field of semiconductor manufacturing technology, and in particular relates to a method and system for automatically generating semiconductor plant layout diagrams based on artificial intelligence models. Background Technology

[0002] In the semiconductor manufacturing industry, generating equipment layout diagrams is crucial for factory operational efficiency. Existing technologies, such as the Zhongxiangying ziFMB system and Visual Factory, provide equipment status monitoring and visualization capabilities. However, these solutions primarily rely on manual drag-and-drop operations to customize layout diagrams on a canvas. This process is complex and time-consuming, requiring users to possess specialized knowledge to manually arrange machines and connect pipes. This manual approach is prone to errors, such as overlooking equipment conflicts or improper connections, and it cannot update the layout in real time to reflect equipment changes, resulting in low accuracy and response delays. Therefore, the main technical problem is that manually generating semiconductor factory layout diagrams is inefficient, error-prone, and unable to achieve real-time and accurate updates. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a method, system, device and medium for automatically generating semiconductor plant layout diagrams based on artificial intelligence models, which significantly reduces the probability of errors in manual operation and improves the overall quality and efficiency of semiconductor plant layout design.

[0004] The first aspect of this invention discloses a method for automatically generating semiconductor plant layout diagrams based on an artificial intelligence model, comprising the following steps:

[0005] Receive semiconductor plant layout data, which includes at least equipment data;

[0006] The received device data is standardized and converted into a unified format;

[0007] The standardized data is input into an artificial intelligence model for vectorization processing to generate numerical vectors representing the characteristics of the equipment. The artificial intelligence model employs a neural network architecture and is trained based on historical semiconductor plant layout data.

[0008] The artificial intelligence model adopts a multilayer perceptron structure, including an input layer, at least one hidden layer, and an output layer. Each neuron in the layer achieves nonlinear feature transformation through an activation function.

[0009] Based on the generated numerical vectors, the optimal arrangement of equipment positions within the semiconductor plant is automatically calculated using a layout optimization algorithm that comprehensively considers equipment physical dimensions, functional correlations, and spatial constraints.

[0010] Based on the arrangement of machine positions, the system automatically generates a pipeline connection scheme between the equipment, including pipeline path planning and connection point location.

[0011] The final semiconductor plant layout diagram is output, which fully presents the distribution of machine locations and pipeline connections in a visual form.

[0012] The above method also includes performing data cleaning processing on the received device data before data standardization processing. The data cleaning processing includes optionally performing data cleaning processing to remove abnormal data, missing data and duplicate data.

[0013] The above method, in the data cleaning step, also includes:

[0014] Statistical analysis methods are used to detect data anomalies. When abnormal data is identified, an error message containing modification suggestions is automatically generated.

[0015] Based on the data modified by the user according to the error message, the data cleaning and standardization process is re-executed.

[0016] The above method, wherein the data standardization processing includes:

[0017] The categorical attributes in the equipment data are numerically encoded.

[0018] Normalize numerical attributes to make their numerical range uniformly between 0 and 1.

[0019] The above method further includes a data validation step before the vectorization processing step:

[0020] The standardized data is compared with the industry rule base, and rule verification is performed on the industry rule base, which includes semiconductor equipment size standards, pipe connection specifications and space constraint rules.

[0021] When the data does not conform to the rules, an error message is generated and returned to the user for modification;

[0022] Only after the data passes verification is it input into the artificial intelligence model for vectorization processing;

[0023] In the vectorization process, the artificial intelligence model is based on a deep neural network architecture and maps the input layout data into numerical vectors through multi-layer nonlinear transformations. Each layer of the neural network includes the calculation process of weight parameters and bias parameters.

[0024] The above method, wherein the layout optimization algorithm achieves machine position optimization by minimizing an objective function, the objective function including a weighted sum of the functional correlation and distance between devices, and a penalty term considering physical constraints.

[0025] In the above method, the physical constraints include equipment safety distance requirements, floor height limits, and equipment size limits, and the penalty term coefficient is configured according to the importance of the constraints.

[0026] The above method, in the step of generating pipe connections, further includes:

[0027] Based on the machine location information and pipeline physical characteristics, a graph search algorithm is used to calculate the optimal pipeline path;

[0028] Conflict detection is performed when determining connection points to avoid pipe crossings and spatial interference.

[0029] The above method also includes a real-time layout update step:

[0030] Continuously monitor changes in equipment layout data;

[0031] When a data update is detected, the entire process from data cleaning to layout generation is automatically triggered, updating the semiconductor plant layout diagram in real time.

[0032] The above method, in the step of outputting the layout diagram, also includes:

[0033] Export layout data in a structured data format to support integration with other semiconductor plant management systems;

[0034] It provides an interactive user interface that allows users to manually adjust and finally confirm the automatically generated layout.

[0035] The second aspect of this invention discloses an automatic semiconductor fab layout generation system based on an artificial intelligence model, comprising:

[0036] The data receiving interface is configured to receive semiconductor plant layout data, which includes at least equipment data.

[0037] A data processing engine, connected to the data receiving interface, is configured to perform data standardization processing on the received device data and convert the data into a unified format.

[0038] The intelligent analysis module, connected to the data processing engine, includes an artificial intelligence model trained based on historical semiconductor plant layout data. The artificial intelligence model adopts a neural network architecture and is configured to input standardized data into the artificial intelligence model for vectorization processing to generate numerical vectors representing equipment characteristics.

[0039] The layout optimization module, connected to the intelligent analysis module, is configured to automatically calculate the optimal position arrangement of the machines in the semiconductor plant based on the generated numerical vector through the layout optimization algorithm. The layout optimization algorithm comprehensively considers the physical size of the equipment, functional correlation and spatial constraints.

[0040] The pipeline connection planning module is connected to the layout optimization module and is configured to automatically generate a pipeline connection scheme between equipment based on the arrangement of machine positions, including pipeline path planning and connection point positioning.

[0041] The visualization output module is connected to the pipeline connection planning module and is configured to output the final semiconductor plant layout diagram, which fully presents the machine location distribution and pipeline connection relationship in a visual form.

[0042] A third aspect of the present invention discloses an electronic device, comprising: a memory and a processor, wherein the processor and the memory are connected;

[0043] The memory is used to store programs;

[0044] The processor invokes a program stored in the memory to execute the method provided by the first aspect embodiment and / or any possible embodiment in combination with the first aspect embodiment.

[0045] The fourth aspect of the present invention discloses a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a computer to perform the methods provided by the embodiments of the first aspect and / or any possible embodiments in combination with the embodiments of the first aspect.

[0046] Compared with existing technologies, this invention has the following advantages: First, by receiving equipment data from multiple sources and performing automated data cleaning and standardization, this invention eliminates common anomalies and repetitive problems encountered in manual operations, providing a reliable data foundation for subsequent processing. Second, by utilizing an artificial intelligence model for vectorization processing, it automatically generates numerical vectors representing equipment characteristics. This step replaces traditional manual analysis, significantly improving data processing speed and reducing layout generation time from hours to minutes. Furthermore, based on the numerical vectors, a layout optimization algorithm automatically calculates the optimal position arrangement of workstations, comprehensively considering equipment physical dimensions, functional relevance, and spatial constraints, avoiding layout conflicts and errors caused by manual dragging. Simultaneously, it automatically generates pipeline connection schemes, including pipeline path planning and connection point positioning, ensuring the accuracy and consistency of connection relationships, supporting real-time data stream input, and enabling rapid response to equipment changes and layout diagram updates. Finally, it outputs a complete layout diagram in a visual form, achieving a fully automated process, significantly reducing the probability of errors in manual operation, and improving the overall quality and efficiency of semiconductor plant layout design.

[0047] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0048] Figure 1This is a flowchart of the method in Embodiment 1 of the present invention.

[0049] Figure 2 This is a system module diagram of Embodiment 2 of the present invention. Detailed Implementation

[0050] Example 1

[0051] like Figure 1 As shown, a method for automatically generating semiconductor plant layout diagrams based on an artificial intelligence model includes the following steps:

[0052] Step 1: Receive semiconductor plant layout data. The layout data includes at least equipment data; for example, it may include equipment data, CAD drawing data, and MES system data, originating from Excel files, databases, or real-time data streams. When the layout also includes CAD drawing data, the server can perform drawing data recognition to analyze the equipment dimensions. When the layout also includes MES system data, the server can perform parsing to analyze changes in equipment location data.

[0053] During implementation, data is received from multiple sources, such as equipment data uploaded by users via Excel files. This includes typical values ​​from the crusher's temperature sensor (30 degrees Celsius within the 20-100 degree Celsius range) and vibration sensor values ​​(2.1 m / s² within the 0-5 m / s² range). Simultaneously, CAD drawing data may be received, such as the cabinet dimensions (3000 mm x 600 mm) in a floor plan, or MES system data, such as the photoresist purity level (99.999%). Data reception is achieved through a file parsing interface or real-time data stream monitoring, ensuring complete data import. This multi-source data integration provides a foundation for subsequent processing, avoids the tedious manual data entry of traditional methods, and improves the efficiency and accuracy of data acquisition.

[0054] Step 2: Standardize the received device data and convert it into a uniform format;

[0055] During implementation, the received data is standardized to a unified format. For example, for the temperature sensor values ​​of the crusher, the Min-Max normalization method is used to convert them to the range of 0 to 1. The calculation method is to subtract the minimum value from the current value and then divide by the maximum value minus the minimum value. Similarly, classification attributes such as equipment type, such as crusher coded as 1 and grinding mill coded as 2, are numerically encoded. The standardization process is implemented through a predefined algorithm module to ensure that all data are scaled consistently. By eliminating dimensional differences, a uniform input is provided to the artificial intelligence model, reducing bias in subsequent processing and improving computational efficiency.

[0056] Step 3: Input the standardized data into the artificial intelligence model for vectorization processing to generate numerical vectors representing the characteristics of the equipment. The artificial intelligence model adopts a neural network architecture. The artificial intelligence model is trained based on historical semiconductor plant layout data. The artificial intelligence model adopts a multilayer perceptron structure, including an input layer, at least one hidden layer and an output layer. Each neuron in the layer realizes nonlinear feature transformation through an activation function.

[0057] In implementation, standardized data is input into an AI model trained on historical semiconductor plant layout data. This model employs a multilayer perceptron neural network architecture. For example, the input layer receives a normalized temperature value of 0.125 and a vibration value of 0.42, assuming a vibration value of 2.1 m / s² is normalized to 0.42. The hidden layers use the ReLU activation function for nonlinear transformation, and the output layer generates a numerical vector such as 0.8, 0.2, and 0.5, representing equipment characteristics such as size, type, and correlation. The training process uses historical data to optimize weight parameters, enabling the model to capture potential patterns in equipment layout. By vectorizing complex data into a machine-understandable format, the intelligence of layout analysis is enhanced, improving processing speed and accuracy.

[0058] Step 4: Based on the generated numerical vectors, the optimal arrangement of the machines in the semiconductor plant is automatically calculated using a layout optimization algorithm. The layout optimization algorithm comprehensively considers the physical size of the equipment, functional correlation, and spatial constraints.

[0059] A layout optimization algorithm is used to calculate equipment positions based on numerical vectors. The algorithm comprehensively considers physical dimensions of the equipment (e.g., a crusher measuring 2 meters by 1 meter), functional interrelationships (e.g., crushers and grinding mills need to be close together to shorten material transport distances), and spatial constraints (e.g., factory area limitations). For example, the objective function minimizes the weighted sum of distances and interrelationships between equipment, and includes penalty terms to handle constraints such as safety clearance requirements. Through iterative optimization, the optimal position arrangement is calculated, such as placing the crusher at coordinates 10 and 20, and the grinding mill at coordinates 15 and 25. By automatically generating a scientifically sound layout, the arbitrariness of manual dragging and dropping is avoided, conflicts and errors are reduced, and layout efficiency is improved.

[0060] Step 5: Based on the machine location arrangement, automatically generate a pipeline connection scheme between the equipment, including pipeline path planning and connection point location;

[0061] During implementation, based on location information, graph search algorithms such as A* algorithm are used to calculate the optimal pipeline path. For example, from the crusher to the grinding mill, the algorithm considers the pipeline's physical characteristics such as diameter and material, plans the path to avoid obstacles, and performs conflict detection to ensure no intersections or spatial interference. Connection point positioning is based on equipment interface coordinates, such as using interface information from CAD drawings. Automated pipeline planning reduces errors and time consumption associated with manual design, ensuring the safety and efficiency of pipeline connections.

[0062] Step 6: Output the final semiconductor plant layout diagram, which presents the complete distribution of machine locations and pipeline connections in a visual form.

[0063] During implementation, a visual layout diagram is generated, presenting the machine locations and piping connections in a graphical interface. For example, the user interface displays icons for crushers and grinders, along with their connecting pipes, supporting zooming and detailed viewing. Output formats include structured data such as JSON for export to other management systems. By providing intuitive visualizations, users can quickly understand and make decisions, improving overall work efficiency.

[0064] In one embodiment, the received device data is further subjected to data cleaning processing before data standardization processing. The data cleaning processing may optionally include performing data cleaning processing to remove abnormal data, missing data, and duplicate data.

[0065] During implementation, the system first checks data integrity. For example, if the vibration sensor data for the crusher in the uploaded Excel file contains missing values, it automatically fills in the default values ​​or deletes the record. Simultaneously, it detects duplicate data; if the same temperature value appears multiple times, it removes duplicates. Optionality means that users can choose whether to perform this step; for example, a checkbox can be provided in the user interface, and the system will only perform cleaning if this option is checked.

[0066] In one embodiment, the data cleaning step further includes: using statistical analysis methods to detect data anomalies; automatically generating error messages containing modification suggestions when abnormal data is identified; and re-executing the data cleaning and standardization process based on the data modified by the user according to the error messages.

[0067] In implementation, statistical analysis methods such as the Z-score algorithm are used to detect data anomalies. For example, when a user uploads temperature sensor data from a crusher via Excel, the system calculates the mean and standard deviation of the data. If the temperature value exceeds the normal range of 20 to 100 degrees Celsius, such as detecting 120 degrees Celsius, it is identified as abnormal data. The system automatically generates an error message, such as: "Temperature value abnormal; it is recommended to check sensor calibration or re-enter the data." The user can modify the data according to the prompt, and then the system re-executes the data cleaning and standardization process to ensure data quality. The above operations improve data reliability, reduce layout errors caused by erroneous data, and thus improve the overall system accuracy.

[0068] In one embodiment, the data standardization process includes: numerically encoding the categorical attributes in the device data; and normalizing the numerical attributes so that their numerical range is uniformly converted to between 0 and 1.

[0069] In implementation, categorical attributes are processed first. For example, the equipment type "crusher" is encoded as the value 1, and "grinding mill" as the value 2, to facilitate processing by the machine learning model. Next, numerical attributes are normalized. For instance, the temperature value of 30 degrees Celsius for the crusher is converted to a range of 0 to 1 using the Min-Max normalization method. Similarly, the current value of 50 amperes is normalized to 0.5. This processing ensures that all data are scaled consistently, eliminating dimensional differences, providing uniform input to the artificial intelligence model, reducing computational bias, and improving processing efficiency.

[0070] In one embodiment, prior to the vectorization process, a data verification step is included: comparing the data against an industry rule base to perform rule verification on the standardized data. The industry rule base includes semiconductor equipment size standards, pipe connection specifications, and spatial constraint rules. When the data does not conform to the rules, an error message is generated and returned to the user for modification. Only when the data passes the verification is it input into the artificial intelligence model for vectorization. In the vectorization process, the artificial intelligence model is based on a deep neural network architecture and maps the input layout diagram data into numerical vectors through multi-layer nonlinear transformations. Each layer of the neural network includes the calculation process of weight parameters and bias parameters.

[0071] During implementation, the system compares against an industry rule base, which includes semiconductor equipment size standards, piping connection specifications, and spatial constraint rules. For example, when processing cabinet size data from CAD drawings, the system verifies whether the dimensions conform to industry standards, such as the cabinet width not exceeding 600 mm. If the data shows a cabinet width of 700 mm, an error message is generated: "Cabinet size exceeds limits, please adjust to within 600 mm." Only after the user makes the correction is the data input into the artificial intelligence model for vectorization processing. Vectorization processing uses a deep neural network architecture, mapping the input data into numerical vectors through multiple layers of nonlinear transformations. For example, an input normalized temperature value of 0.125, after passing through hidden layers to calculate weights and bias parameters, output vectors such as 0.8, 0.2, and 0.5.

[0072] In one embodiment, the layout optimization algorithm optimizes machine positions by minimizing an objective function, which includes a weighted sum of functional correlation and distance between devices, as well as a penalty term that considers physical constraints.

[0073] It should be noted that in the implementation, the correlation between equipment is calculated. For example, due to material transfer requirements, the correlation weight between the crusher and the grinding mill is set to 0.8, and the distance weight is set to 0.2. The objective function minimizes the weighted sum of the total distance and correlation, while incorporating a penalty term to handle constraints. For instance, if the distance between equipment is less than the safety requirement of 1 meter, a penalty value is added. The optimal positions are calculated using iterative optimization algorithms such as gradient descent, for example, placing the crusher at coordinates 10, 20, and the grinding mill at coordinates 15, 25.

[0074] In one embodiment, the physical constraints include equipment safety distance requirements, floor height limits, and equipment size limits, with penalty coefficients configured according to the importance of the constraints.

[0075] In implementation, for example, if the safety distance requirement is set to 1 meter, and the algorithm calculates a device spacing of less than 1 meter, the penalty coefficient is configured to 0.5 based on importance to increase the penalty. Floor height limits, such as not exceeding 3 meters, and equipment size limits, such as a crusher length not exceeding 2 meters, are also set, and these constraints are checked in real time during the optimization process.

[0076] In one embodiment, the step of generating pipe connections further includes: calculating the optimal pipe path using a graph search algorithm based on machine location information and pipe physical characteristics; and performing conflict detection when determining connection points to avoid pipe crossings and spatial interference.

[0077] In implementation, for example, the crusher is located at coordinates 10,20, the grinding mill at coordinates 15,25, and the selection equipment at other specific locations. This location information comes from calculations performed using previous layout optimization algorithms. Physical characteristic data of these devices is obtained, including pipe diameter, material type, and connection interface specifications. This data may come from CAD drawings or device sensor data. For example, CAD drawings may specify the use of ZR-YJV-5x6 cables or copper busbar paths, while device sensors provide real-time parameters such as a flow rate of 500 cubic meters per hour or a pressure of 5 bar.

[0078] Next, a graph search algorithm is used to calculate the optimal path for the pipeline. Specifically, the A* algorithm is used, a highly efficient path planning method capable of finding the shortest or optimal path under various constraints. The algorithm models the entire semiconductor plant space as a graph structure, where nodes represent possible locations, edges represent pipeline segments, and each edge has a weight based on distance, pipeline cost, or other factors such as construction difficulty. For example, in path planning from the crusher to the grinding mill, the algorithm starts from the connection interface of the crusher and progressively explores adjacent nodes, evaluating each possible path option. The evaluation function includes actual movement costs (such as pipeline length) and heuristically estimated costs (such as the straight-line distance to the destination) to ensure that the path is both short and feasible. In this process, the system considers the physical characteristics of the pipeline; for example, if the pipeline material is copper, its bending radius limitation may affect the path selection, and the algorithm avoids overly sharp turns to meet engineering standards. At the same time, spatial constraints such as plant structure and the location of other equipment are also incorporated into the graph model as obstacle nodes, and the algorithm automatically avoids these areas to ensure that the path does not cross physical obstacles.

[0079] While calculating the path, conflict detection is performed to avoid pipe crossings and spatial interference. The conflict detection module, based on real-time spatial analysis, uses geometric algorithms to check if the pipe path intersects with existing structures or other pipes. For example, when planning a pipe from a crusher to a grinding mill, the system simulates the pipe's path in three-dimensional space, checking if it intersects with cable trays, ventilation ducts, or cabinet clearance areas on a floor. If a potential conflict is detected, such as the pipe path overlapping the coordinates of a UPS distribution cabinet, the problem is immediately flagged, and the path is automatically adjusted or a warning is issued. Adjustments may involve slight path offsets, selecting alternative routes, or prompting user intervention, such as highlighting the conflict point on the user interface and suggesting modifications to the pipe route or equipment location. This detection not only considers static elements but also dynamically responds to real-time data changes; for example, if the MES system updates equipment locations, conflict detection will rerun to ensure continuous compliance.

[0080] In one embodiment, the method further includes a real-time layout update step: continuously monitoring changes in the equipment layout data; and automatically triggering a full-process handling procedure from data cleaning to layout generation when a data update is detected, thereby updating the semiconductor plant layout in real time.

[0081] In practice, for example, if the photoresist data in the MES system is updated, the change is detected through a monitoring mechanism, and then the process is automatically restarted from data cleaning to layout generation, updating the semiconductor plant layout diagram in real time.

[0082] In one embodiment, the step of outputting the layout diagram further includes: exporting the layout diagram data in a structured data format to support integration with other semiconductor plant management systems; and providing an interactive user interface to allow users to manually adjust and finally confirm the automatically generated layout diagram.

[0083] During implementation, for example, after generating the layout diagram, it can be exported as JSON or XML format for easy import into the ERP system. The user interface allows users to drag and drop the positions of the equipment, such as moving the crusher icon. The system recalculates the pipeline connections in real time, and saves the data after user confirmation.

[0084] Example 2

[0085] like Figure 2 As shown, an automatic semiconductor plant layout diagram generation system based on an artificial intelligence model includes:

[0086] The data receiving interface is configured to receive semiconductor plant layout data, which includes equipment data, CAD drawing data, and MES system data, and originates from Excel files, databases, or real-time data streams.

[0087] A data processing engine, connected to the data receiving interface, is configured to perform data standardization processing on the received device data and convert the data into a unified format.

[0088] The intelligent analysis module, connected to the data processing engine, includes an artificial intelligence model trained based on historical semiconductor plant layout data. The artificial intelligence model adopts a neural network architecture and is configured to input standardized data into the artificial intelligence model for vectorization processing to generate numerical vectors representing equipment characteristics.

[0089] The layout optimization module, connected to the intelligent analysis module, is configured to automatically calculate the optimal position arrangement of the machines in the semiconductor plant based on the generated numerical vector through the layout optimization algorithm. The layout optimization algorithm comprehensively considers the physical size of the equipment, functional correlation and spatial constraints.

[0090] The pipeline connection planning module is connected to the layout optimization module and is configured to automatically generate a pipeline connection scheme between equipment based on the arrangement of machine positions, including pipeline path planning and connection point positioning.

[0091] The visualization output module is connected to the pipeline connection planning module and is configured to output the final semiconductor plant layout diagram, which fully presents the machine location distribution and pipeline connection relationship in a visual form.

[0092] The semiconductor plant layout diagram automatic generation system based on artificial intelligence model provided in this embodiment has the same implementation principle and technical effect as the method embodiment in Embodiment 1. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in Embodiment 1.

[0093] Example 3

[0094] A computer-readable storage medium having a computer program stored thereon, the computer program being executed by a computer as described in Embodiment 1 above, which is an automatic generation method for semiconductor plant layout diagrams based on an artificial intelligence model.

[0095] Example 4

[0096] An electronic device includes: a memory and a processor, wherein the processor and the memory are connected;

[0097] The memory is used to store programs;

[0098] The processor calls a program stored in the memory to execute a method for automatically generating semiconductor plant layout diagrams based on an artificial intelligence model, as described in Embodiment 1.

[0099] It should be noted that the electronic device mentioned may be, but is not limited to, personal computers (PCs), tablet computers, mobile internet devices (MIDs), etc.

[0100] It should be noted that processors, memory, and other components that may be present in electronic devices are electrically connected to each other, directly or indirectly, to enable data transmission or interaction. For example, processors, memory, and other components may be electrically connected to each other via one or more communication buses or signal lines.

[0101] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0102] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0103] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0104] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, laptop, server, mobile phone, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the present invention. Any simple modifications, alterations, or equivalent structural changes made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for automatically generating semiconductor plant layout diagrams based on an artificial intelligence model, characterized in that, Includes the following steps: Receive semiconductor plant layout data, which includes at least equipment data; Before standardizing the received device data, data cleaning is performed. This data cleaning process may include optionally removing outlier, missing, and duplicate data. The received device data is standardized and converted into a unified format; The standardized data is input into an artificial intelligence model for vectorization processing to generate numerical vectors representing equipment characteristics. This artificial intelligence model employs a neural network architecture and is trained based on historical semiconductor plant layout data. Before vectorization, the process includes: comparing the standardized data against an industry rule base to perform rule validation. This industry rule base includes semiconductor equipment size standards, pipe connection specifications, and spatial constraint rules. If the data does not conform to the rules, an error message is generated and returned to the user for modification. Only data that passes validation is input into the artificial intelligence model for vectorization processing. Based on the generated numerical vectors, the optimal arrangement of equipment positions within the semiconductor plant is automatically calculated using a layout optimization algorithm that comprehensively considers equipment physical dimensions, functional correlations, and spatial constraints. The layout optimization algorithm optimizes machine positions by minimizing an objective function. This objective function includes a weighted sum of functional interrelationships and distances between machines, as well as a penalty term considering physical constraints. These physical constraints include machine safety distance requirements, floor height limits, and machine size limitations. The penalty term coefficient is configured according to the importance of the constraints. Based on the arrangement of machine positions, the system automatically generates a pipeline connection scheme between the equipment, including pipeline path planning and connection point location. Based on machine location information and pipeline physical characteristics, a graph search algorithm is used to calculate the optimal pipeline path; conflict detection is performed when determining connection points to avoid pipeline crossings and spatial interference. Continuously monitor changes in equipment layout data; when data updates are detected, automatically trigger the entire process from data cleaning to layout generation, and update the semiconductor plant layout in real time; The final semiconductor plant layout diagram is output, which fully presents the distribution of machine locations and pipeline connections in a visual form.

2. The method according to claim 1, characterized in that, The data cleaning process also includes: Statistical analysis methods are used to detect data anomalies. When abnormal data is identified, an error message containing modification suggestions is automatically generated. Based on the data modified by the user according to the error message, the data cleaning and standardization process is re-executed.

3. The method according to claim 1, characterized in that, The data standardization process includes: The categorical attributes in the equipment data are numerically encoded. Normalize numerical attributes to make their numerical range uniformly between 0 and 1.

4. A semiconductor plant layout diagram automatic generation system based on an artificial intelligence model, characterized in that, include: The data receiving interface is configured to receive semiconductor plant layout data, which includes at least equipment data. Before standardizing the received equipment data, data cleaning is performed, optionally removing outlier, missing, and duplicate data. A data processing engine, connected to the data receiving interface, is configured to perform data standardization processing on the received device data and convert the data into a unified format. The intelligent analysis module, connected to the data processing engine, includes an artificial intelligence model trained based on historical semiconductor plant layout data. This model employs a neural network architecture and is configured to input standardized data into the AI ​​model for vectorization processing, generating numerical vectors representing equipment characteristics. Before vectorization, the module further includes: comparing the standardized data against an industry rule base to perform rule validation. The industry rule base includes semiconductor equipment size standards, pipe connection specifications, and spatial constraint rules. When data does not conform to the rules, an error message is generated and returned to the user for modification. Only data that passes validation is input into the AI ​​model for vectorization processing. A layout optimization module, connected to the intelligent analysis module, is configured to automatically calculate the optimal arrangement of machines within the semiconductor factory based on generated numerical vectors using a layout optimization algorithm. This algorithm comprehensively considers equipment physical dimensions, functional interrelationships, and spatial constraints. The algorithm optimizes machine placement by minimizing an objective function, which includes a weighted sum of functional interrelationships and distances between machines, as well as a penalty term considering physical constraints. These physical constraints include equipment safety distance requirements, floor height limits, and equipment size limitations. The penalty term coefficient is configured according to the importance of the constraints. The pipeline connection planning module, connected to the layout optimization module, is configured to automatically generate pipeline connection schemes between equipment based on the machine position arrangement, including pipeline path planning and connection point location; calculate the optimal pipeline path using a graph search algorithm based on machine position information and pipeline physical characteristics; and perform conflict detection when determining connection points to avoid pipeline crossings and spatial interference. It also includes a real-time layout update step: continuously monitoring changes in equipment layout data; when a data update is detected, automatically triggering the entire process from data cleaning to layout generation, and updating the semiconductor plant layout in real time; The visualization output module is connected to the pipeline connection planning module and is configured to output the final semiconductor plant layout diagram, which fully presents the machine location distribution and pipeline connection relationship in a visual form.

5. An electronic device, characterized in that, include: A memory and a processor, wherein the processor and the memory are connected; The memory is used to store programs; The processor invokes a program stored in the memory to execute the method as described in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, It stores a computer program, which is executed by a computer to perform the method as described in any one of claims 1-3.

Images