An Artificial Intelligence-Based Integrated Circuit Manufacturing Monitoring System and Method
By using artificial intelligence monitoring systems and methods, the problems of long production cycles and high losses in integrated circuit manufacturing have been solved, achieving efficient and precise production chain management and improving yield and production efficiency.
Patent Information
- Application Number
- CN202310054530.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-02-03
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-02-03
AI Technical Summary
In the current integrated circuit manufacturing process, independent processing by each manufacturer leads to longer production cycles, lower yield rates, significant losses during transportation, and a loosely connected production chain, making it difficult to achieve efficient production.
An AI-based integrated circuit production monitoring system is adopted, including a processing monitoring module, a scheduling module, and an inbound/outbound monitoring module. It monitors the processing steps through Boolean signals, avoids omissions through voltage monitoring, optimizes the processing sequence and transportation losses, and achieves global scheduling and replenishment.
Shorten production cycle, increase yield, reduce transportation losses, optimize production chain connections, improve production efficiency, and ensure the accuracy and quality of integrated circuit processing.
Smart Images

Figure CN116895543B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of integrated circuit manufacturing, specifically to an integrated circuit manufacturing monitoring system and method based on artificial intelligence. Background Technology
[0002] Artificial intelligence is an emerging technology discipline based on computer science, integrating multiple disciplines, and simulating human intelligence. Its research areas include robotics, speech recognition, image recognition, and expert systems. An integrated circuit is a miniature electronic device or component that interconnects the components and wiring required for a circuit, fabricates them on a small semiconductor chip, and then packages them to become a miniature structure with circuit functions. The development of integrated circuits has greatly improved electronic components in terms of miniaturization, intelligence, and high reliability. The current direction of integrated circuit production is the separation of design and manufacturing. In manufacturing, existing integrated circuits are mainly based on silicon wafers, with some based on selenium or other materials. The current mainstream integrated circuits are made into large-size wafers from silicon, and production is completed by processing, manufacturing, packaging and testing the wafers. Existing integrated circuit manufacturing is typically carried out by large, qualified enterprises, and the upstream and downstream supply chains are complex. Each processing step needs to be completed independently by different manufacturers, with unclear completion times and quality. The components are then transferred to the next factory for further processing. The processing and waiting times vary between manufacturers, leading to longer production cycles. At the same time, components are damaged during the transfer process, resulting in components that do not meet processing conditions when reprocessed, reducing yield and delaying the schedule. This paper proposes an integrated circuit production monitoring system and method based on artificial intelligence to solve the above problems. Summary of the Invention
[0003] The purpose of this invention is to provide an integrated circuit production monitoring system and method based on artificial intelligence to solve the problems mentioned in the background art.
[0004] To achieve the above objectives, the present invention provides the following technical method: an integrated circuit production monitoring system and method based on artificial intelligence, characterized in that: the production monitoring system includes a processing monitoring module, a scheduling module, and an inbound / outbound monitoring module; The processing monitoring module monitors each processing step according to the production process, obtains and corrects the real-time processing status of the work line, and predicts the working time required for each processing step to complete its processing task based on the production line efficiency. The scheduling module receives the processing status of the production line and adjusts the processing procedures within the factory through a scheduling algorithm, integrating and connecting manufacturers across the entire production chain to optimize the processing sequence between different manufacturers. The inbound and outbound monitoring module monitors detailed information on finished products processed and shipped from the factory, and performs inbound inspection after being transferred to the next processing stage factory to monitor transportation losses. The above technical solutions can integrate the production chain, optimize industrial processing connections, reduce waiting time between different factories, improve the tightness of upstream and downstream connections in the production chain, shorten the production cycle, and improve production efficiency; conduct inbound and outbound inspections in the processing stage, obtain loss data generated in the transportation stage through inspection, optimize transportation and packaging choices, reduce losses caused during transit, and replenish purchases based on loss data to meet production needs. According to the above technical solution, the processing monitoring module includes a process monitoring unit, a voltage monitoring unit, a correction unit, and a prediction unit; The process monitoring unit monitors the integrated circuit manufacturing process steps. It acquires data from the processing machinery via sensors electrically connected to the machinery and checks whether the integrated circuit has completed its processing action after each step. Due to the complexity and lengthy processes of integrated circuit manufacturing, and the periodic errors in machinery operation, some integrated circuits may not be fully processed. This unit monitors all integrated circuits to prevent omissions in processing steps, thus avoiding a high rate of defective products. The voltage monitoring unit monitors the voltage status of the production line. The correction unit issues an alarm for abnormal temperatures on the production line and performs secondary processing to repair integrated circuits that have missed processing steps before they enter the next production step. The lead time prediction unit estimates the completion time based on the current production line efficiency and the production task of this batch of integrated circuits, making the entire integrated circuit production process quantifiable and controllable. According to the above technical solution, the scheduling module includes an information receiving unit, a communication unit, a process adjustment unit, and a coordination unit; The information receiving unit receives information about processing progress from different manufacturers, providing basic data for system coordination; the communication unit connects upstream and downstream manufacturers, providing communication services and a platform for information exchange, reducing information barriers; the process adjustment unit adjusts the processing quantity and time based on the workload of each processing stage in the factory, and rationally arranges the processing plan; the coordination unit optimizes the processing sequence based on the production time between manufacturers, reducing idle waiting time between manufacturers. According to the above technical solution, the inbound and outbound monitoring module includes an outbound information unit, an inbound information unit, a transportation monitoring unit, and a loss unit; After receiving the scheduling task, the outbound information unit records the outbound information of integrated circuit parts. In the transportation monitoring unit, the location of the integrated circuit transportation and the estimated arrival time are located. The loss unit monitors the loss caused during packaging, transportation and handling, and submits a supplementary application for integrated circuit parts according to the estimated delivery quantity. The inbound information unit records the received integrated circuit parts information and records the proportion of supplementary integrated circuit parts and transported integrated circuit components. An artificial intelligence-based method for monitoring integrated circuit production, comprising the following steps: S1. Monitor the processing status of the production line, monitor the completion status of each part processing step, and correct any missing or incomplete steps; S2. Monitor voltage fluctuations during production line processing and predict the production period based on real-time data; S3. Receive production information, schedule the work sequence of the production line within the factory, integrate the upstream and downstream of the production chain, and coordinate the scheduling between different manufacturers. S4. Monitor the inbound and outbound status after scheduling, and monitor transportation time and losses during transportation; According to the above technical solution, in step S1, it is monitored whether the part processing step is missing. Integrated circuit production involves numerous intermediate steps and batch processing. Through mechanical and chemical processing, random errors inevitably occur during the process. Simultaneously, mechanical failures and step errors can also cause individual production steps to fail to complete on schedule. To monitor the completion status of each processing step, a method is proposed: S101. In the initial processing stage, a Boolean signal is placed in this stage, and the Boolean signal contains only two data: 0 and 1. S102. In each processing stage, after a processing action occurs, a Boolean signal is incremented once, and a Boolean signal is sent to the next processing stage. The Boolean signal recorded in this stage is 0, and it increments to 1 after the processing action. S103. Before the next processing stage begins, detect the Boolean signal. If the Boolean signal is 1, it means that the integrated circuit was processed in the previous processing stage, and the current processing stage continues. If the Boolean signal is 0, it means that there was a problem with the processing of the integrated circuit in the previous processing stage. At this time, the integrated circuit is transported through another channel and transferred back to the entry end of the previous processing stage. The reverse is also true for the Boolean signal, which only jumps between 0 and 1. Without changing the order of the processing stages, the Boolean signal received in each processing stage remains unchanged. Each integrated circuit has its own Boolean signal sequence. Based on this signal jumping mechanism, after the normal processing stage is completed, the adjacent Boolean signals must be different. Plotting the Boolean signal sequence into a graph shows periodicity, and the production processing status can be obtained through each Boolean signal sequence graph. According to the above technical solution, in step S2, the processing stage is electrically driven. Since the integrated circuit processing is a precision operation with high accuracy requirements, the power quality changes caused by voltage fluctuations will affect the life of the equipment motor and cause unstable motor speed, which will adversely affect the precision process of integrated circuits. Now, the voltage monitoring stage is used to monitor the real-time voltage of the processing stage. In the power supply, after the transformer conversion, the rated voltage is continuously input. The input voltage is generally stable within a time period. Based on the multi-functional instrument, it is detected whether the voltage meets the rated voltage requirements of the processing machinery, and alarm processing is carried out for voltage drop and overvoltage phenomena. The input voltage fluctuates at a certain moment, which is difficult to monitor by instruments, but it will also affect the equipment. Now, the voltage flicker measurement method based on FFT is used to monitor the instantaneous voltage flicker value and record the batch number of the integrated circuit produced at the flicker moment for inspection to see if the instantaneous voltage flicker affects the processing. The instantaneous flicker value is monitored and calculated through the following steps. S201: Using the production time of each batch as the time period, calculate the root mean square voltage sequence U(t) within the period; sample the voltage signal u(t) at n points within each half period to obtain the sampled signal u(n), and calculate a root mean square voltage value for u(n) for each half period; S202: Transform the sequence U(t) to obtain its spectrum; the definition of voltage fluctuation is: ; Where u max and u min U represents two adjacent extreme values of the root mean square voltage. t Given the rated voltage; perform a Fast Fourier Transform on the sequence U(t) to obtain the discrete spectrum, and divide twice the amplitude of the spectrum at the corresponding frequency by the amplitude of the decomposed DC component to obtain the voltage fluctuation D(i). Calculate the instantaneous flicker value at that frequency: ; Where σ is the voltage fluctuation required to generate a unit instantaneous flicker value at this frequency, and this value can be obtained by fitting it from the IEC61000-4-15 standard; S203: Calculate the instantaneous flicker value corresponding to the root mean square voltage sequence U(t); sum the instantaneous flicker values corresponding to each frequency, and obtain the instantaneous flicker value corresponding to U(t). ; S204: Calculate short-time flicker value; Short-time flicker value: ; Generally, ρ1 is 0.0314, ρ2 is 0.0525, ρ3 is 0.0657, ρ4 is 0.28, ρ5 is 0.08, and Q... i Q j Q k Q l Q m These are the instantaneous flicker values for 0.1%, 1%, 3%, 10%, and 50% of the time within the period, respectively. For other conditions, values can be obtained from the fitted data. After obtaining Q s Next, it is determined whether the instantaneous flicker value of the current processing stage exceeds the limit range. This limit range can be obtained by monitoring multiple production data and the correlation with product quality. Processing batches with instantaneous flicker values exceeding the limit range are re-inspected to avoid integrated circuit quality problems. In the prediction unit, the production unit integrated circuit time is based on the production process. and production plan Calculate the expected production completion time ; According to the above technical solution, in step S3, the production process and transportation are further optimized based on real-time data and project schedule prediction results; some production processes of integrated circuits have a sequential order, and some production processes can be scheduled within the scope, and the scheduling method is as follows; S301. Set the processing stage to W, the processing stage state to Wi, and the processing stage states to busy and idle respectively. The processing stage W is activated under the idle condition. S302. Retrieve integrated circuit batches that have completed the production process and are in the waiting stage; S303. Search all processing steps. Add processing steps with a fixed sequence to an ordered sequence of strongly constrained processing steps. Mark the processing steps of each batch of integrated circuits. Mark the completed processing steps and search for the processing steps that have not yet been performed. At this time, the processing steps before the idle processing steps are all busy or marked. Find the idle Wi and determine whether the processing needs to be transferred. If it needs to be transferred, determine the waiting time and the transfer processing time respectively. Based on the comparison results, submit a partial transfer application to the downstream factory to process and produce part of the batch of integrated circuits in advance. If there are no idle processing steps, jump to S304. S304. For processing steps with a fixed sequence, regardless of whether they are busy or idle, set them to busy and deactivate them. For interchangeable steps, set them to busy or idle according to their actual state and add them all to an ordered sequence of processing steps based on weak constraints. At the same time, mark the processing steps that have been completed for each batch of integrated circuits. Set the search to n steps forward each time. At this time, the processing steps before the idle processing steps are all busy or marked. Select the idle processing steps that do not need to be transferred for processing. If there are no idle processing steps, the search ends. The number of integrated circuits to be transferred to the processing steps is determined by the working efficiency of the work steps that are about to be transferred. Usually, it is set that the production of some transferred integrated circuits is completed before the next production cycle. The purpose of this step is to make the most of idle processing stages and reduce the time required for production. According to the above technical solution, in step S4, based on scheduling requirements, the inbound and outbound status is recorded, the transportation process is monitored, the transportation arrival time is predicted, and the losses caused during the transportation process are monitored. After the scheduling task is issued, the outbound unit receives the scheduling task order, disassembles the data to obtain the transfer destination and transfer quantity, and issues a transfer outbound task order, recording the outbound data. In the integrated circuit outbound transportation process, the transportation mode is monitored, and the transfer vehicle is accurately located based on the positioning system to obtain accurate location information and estimate the transfer arrival time. During the transportation process, if losses occur due to unexpected accidents, the loss unit promptly captures and uploads the data to the system so that other means can be used to quickly replenish the lost portion. When the next manufacturer conducts inbound inventory, the outbound unit data and the loss unit data are connected to achieve precise transportation.
[0005] Compared with the prior art, the beneficial effects achieved by the present invention are: This invention collects information from the production process and, considering the intricate and complex nature of integrated circuit manufacturing, establishes a feasible signal conversion prompt mechanism to prevent omissions in the processing steps that could lead to production errors. For electrically driven processing steps, a voltage monitoring method is introduced to monitor instantaneous voltage flicker and mark specific batches of integrated circuits at the moment of voltage flicker. Batches with flicker values exceeding a certain limit are detected to prevent integrated circuit production failures caused by mechanical disturbances. This invention integrates the entire industry chain, taking a global perspective and setting constraints to cyclically search the processing steps, making the most of idle processing steps to reduce production time. After scheduling, the invention monitors the warehousing and transportation processes to obtain information on losses in real time and to replenish integrated circuits accordingly. Attached Figure Description
[0006] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the module composition of an integrated circuit production monitoring system based on artificial intelligence according to the present invention; Figure 2 This is a flowchart illustrating the steps of an artificial intelligence-based integrated circuit production monitoring method. Detailed Implementation
[0007] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0008] like Figures 1-2 As shown, the present invention provides a technical solution: an integrated circuit production monitoring system and method based on artificial intelligence, the production monitoring system including a processing monitoring module, a scheduling module and an inbound / outbound monitoring module; The processing monitoring module monitors each processing stage according to the production steps, obtains and corrects the real-time processing status of the workline, and predicts the working time required for each processing stage to complete its processing task based on the production line efficiency. The scheduling module receives the processing status of the production line and adjusts the processing procedures within the factory through a scheduling algorithm, integrating and connecting manufacturers across the entire production chain and optimizing the processing sequence between different manufacturers. The inbound / outbound monitoring module monitors detailed information on finished products processed and shipped from the factory, and performs inbound inspection after being transferred to the next processing stage factory, monitoring transportation losses. This achieves the integration of the production chain, optimizes industrial processing connections, reduces waiting time between different factories, improves the tightness of the upstream and downstream connections in the production chain, shortens the production cycle, and improves production efficiency. By detecting inbound and outbound processes at the processing stages, it obtains loss data generated during transportation to optimize transportation and packaging choices, reduce losses during transportation, and replenish supplies based on loss data to meet production needs. like Figure 1 As shown, the processing monitoring module includes a process monitoring unit, a voltage monitoring unit, a correction unit, and a prediction unit; The process monitoring unit monitors the integrated circuit manufacturing process steps. It acquires data from the processing machinery via sensors electrically connected to the machinery and checks whether the integrated circuit has completed its processing action after each step. Due to the complexity and lengthy processes of integrated circuit manufacturing, and the periodic errors in machinery operation, some integrated circuits may not be fully processed. This unit monitors all integrated circuits to prevent omissions in processing steps, thus avoiding a high rate of defective products. The voltage monitoring unit monitors the voltage status of the production line. The correction unit issues an alarm for abnormal temperatures on the production line and performs secondary processing to repair integrated circuits that have missed processing steps before they enter the next production step. The lead time prediction unit estimates the completion time based on the current production line efficiency and the production task of this batch of integrated circuits, making the entire integrated circuit production process quantifiable and controllable. The scheduling module includes an information receiving unit, a communication unit, a process adjustment unit, and a coordination unit; The information receiving unit receives information about processing progress from different manufacturers, providing basic data for system coordination; the communication unit connects upstream and downstream manufacturers, providing communication services and a platform for information exchange, reducing information barriers; the process adjustment unit adjusts the processing quantity and time based on the workload of each processing stage in the factory, and rationally arranges the processing plan; the coordination unit optimizes the processing sequence based on the production time between manufacturers, reducing idle waiting time between manufacturers. The inbound and outbound monitoring module includes an outbound information unit, an inbound information unit, a transportation monitoring unit, and a loss unit; After receiving the scheduling task, the outbound information unit records the outbound information of integrated circuit parts. In the transportation monitoring unit, the location of the integrated circuit transportation and the estimated arrival time are located. The loss unit monitors the loss caused during packaging, transportation and handling, and submits a supplementary application for integrated circuit parts according to the estimated delivery quantity. The inbound information unit records the received integrated circuit parts information and records the proportion of supplementary integrated circuit parts and transported integrated circuit components. An artificial intelligence-based method for monitoring integrated circuit production, the method as follows: Figure 2 As shown, it includes the following steps: S1. Monitor the processing status of the production line, monitor the completion status of each part processing step, and correct any missing steps; in step S1, monitor whether any part processing steps are missing; integrated circuit production involves numerous intermediate steps and batch processing, using mechanical and chemical processing. Random errors are unavoidable during the processing, and mechanical failures and step errors can also cause some production steps to fail to complete on schedule. To monitor the completion status of each processing step, a method is proposed: S101. In the initial processing stage, a Boolean signal is placed in this stage, and the Boolean signal contains only two data: 0 and 1. S102. In each processing stage, after a processing action occurs, a Boolean signal is incremented once, and a Boolean signal is sent to the next processing stage. The Boolean signal recorded in this stage is 0, and it increments to 1 after the processing action. S103. Before the next processing stage begins, detect the Boolean signal. If the Boolean signal is 1, it means that the integrated circuit was processed in the previous processing stage, and the current processing stage continues. If the Boolean signal is 0, it means that there was a problem with the processing of the integrated circuit in the previous processing stage. At this time, the integrated circuit is transported through another channel and transferred back to the entry end of the previous processing stage. The reverse is also true for the Boolean signal, which only jumps between 0 and 1. Without changing the order of the processing stages, the Boolean signal received in each processing stage remains unchanged. Each integrated circuit has its own Boolean signal sequence. Based on this signal jumping mechanism, after the normal processing stage is completed, the adjacent Boolean signals must be different. Plotting the Boolean signal sequence into a graph shows periodicity, and the production processing status can be obtained through each Boolean signal sequence graph. S2. Monitor voltage fluctuations during production line processing and predict the production period based on real-time data. In step S2, the processing stage is electrically driven. Since the integrated circuit processing is a precision operation with high accuracy requirements, the power quality changes caused by voltage fluctuations will affect the life of the equipment motor and cause unstable motor speed, which will adversely affect the precision process of integrated circuits. Now, the voltage monitoring stage is used to monitor the real-time voltage of the processing stage. In the power supply, after the transformer conversion, the rated voltage is continuously input. The input voltage is generally stable within a time period. Based on the multi-functional instrument, it is detected whether the voltage meets the rated voltage requirements of the processing machinery, and alarm processing is carried out for voltage drop and overvoltage phenomena. The input voltage fluctuates at a certain moment, which is difficult to monitor by instruments, but it will also affect the equipment. Now, the voltage flicker measurement method based on FFT is used to monitor the instantaneous voltage flicker value and record the batch number of the integrated circuit produced at the flicker moment for inspection to see if the instantaneous voltage flicker affects the processing. The instantaneous flicker value is monitored and calculated through the following steps. S201: Using the production time of each batch as the time period, calculate the root mean square voltage sequence U(t) within the period; sample the voltage signal u(t) at n points within each half period to obtain the sampled signal u(n), and calculate a root mean square voltage value for u(n) for each half period; S202: Transform the sequence U(t) to obtain its spectrum; the definition of voltage fluctuation is: ; Where u max and u min U represents two adjacent extreme values of the root mean square voltage. t Given the rated voltage; perform a Fast Fourier Transform on the sequence U(t) to obtain the discrete spectrum, and divide twice the amplitude of the spectrum at the corresponding frequency by the amplitude of the decomposed DC component to obtain the voltage fluctuation D(i). Calculate the instantaneous flicker value at that frequency. ; Where σ is the voltage fluctuation required to generate a unit instantaneous flicker value at this frequency, and this value can be obtained by fitting it from the IEC61000-4-15 standard; S203: Calculate the instantaneous flicker value corresponding to the root mean square voltage sequence U(t); sum the instantaneous flicker values corresponding to each frequency, and obtain the instantaneous flicker value corresponding to U(t). ; S204: Calculate short-time flicker value; Short-time flicker value: ; Generally, ρ1 is 0.0314, ρ2 is 0.0525, ρ3 is 0.0657, ρ4 is 0.28, ρ5 is 0.08, and Q...i Q j Q k Q l Q m These are the instantaneous flicker values for 0.1%, 1%, 3%, 10%, and 50% of the time within the period, respectively. For other conditions, values can be obtained from the fitted data. After obtaining Q s Next, it is determined whether the instantaneous flicker value of the current processing stage exceeds the limit range. This limit range can be obtained by monitoring multiple production data and the correlation with product quality. Processing batches with instantaneous flicker values exceeding the limit range are re-inspected to avoid integrated circuit quality problems. In the prediction unit, the production unit integrated circuit time is based on the production process. and production plan Calculate the expected production completion time ; S3. Receive production information, schedule the workflow of the production line within the factory, integrate the upstream and downstream of the production chain, and coordinate the scheduling between different manufacturers; in step S3, based on real-time data and schedule prediction results, further optimize the production process and transportation; some production processes of integrated circuits have a sequential order, and some production processes can be scheduled within a certain range, and the scheduling method is as follows; S301. Set the processing stage to W, the processing stage state to Wi, and the processing stage states to busy and idle respectively. The processing stage W is activated under the idle condition. S302. Retrieve integrated circuit batches that have completed the production process and are in the waiting stage; S303. Search all processing steps. Add processing steps with a fixed sequence to an ordered sequence of strongly constrained processing steps. Mark the processing steps of each batch of integrated circuits. Mark the completed processing steps and search for the processing steps that have not yet been performed. At this time, the processing steps before the idle processing steps are all busy or marked. Find the idle Wi and determine whether the processing needs to be transferred. If it needs to be transferred, determine the waiting time and the transfer processing time respectively. Based on the comparison results, submit a partial transfer application to the downstream factory to process and produce part of the batch of integrated circuits in advance. If there are no idle processing steps, jump to S304. S304. For processing steps with a fixed sequence, regardless of whether they are busy or idle, set them to busy and deactivate them. For interchangeable steps, set them to busy or idle according to their actual state and add them all to an ordered sequence of processing steps based on weak constraints. At the same time, mark the processing steps that have been completed for each batch of integrated circuits. Set the search to n steps forward each time. At this time, the processing steps before the idle processing steps are all busy or marked. Select the idle processing steps that do not need to be transferred for processing. If there are no idle processing steps, the search ends. The number of integrated circuits to be transferred to the processing steps is determined by the working efficiency of the work steps that are about to be transferred. Usually, it is set that the production of some transferred integrated circuits is completed before the next production cycle. S4. Monitor the inbound and outbound status after scheduling, and monitor transportation time and losses during transportation. In step S4, based on scheduling requirements, record the inbound and outbound status, monitor the transportation process, predict transportation arrival time, and monitor losses caused during transportation. After the scheduling task is issued, the outbound unit receives the scheduling task order, disassembles the data to obtain the transfer destination and transfer quantity, and issues a transfer outbound task order, records the outbound data, monitors the transportation mode in the integrated circuit outbound transportation process, and performs precise positioning of the transfer vehicle based on the positioning system to obtain accurate location information and estimate the transfer arrival time. During transportation, if losses occur due to unexpected accidents, the loss unit promptly captures and uploads the data to the system so that it can be quickly dispatched from other sources to replenish the lost portion. When the next manufacturer performs inbound inventory, the outbound unit data and the loss unit data are connected to achieve precise transportation. Example 1
[0009] If a certain batch of production time is 2s, with 120V and 3.5HZ, calculate the root mean square voltage sequence U(t) within the period; sample the voltage signal u(t) at 6400 points within 1s of each half period to obtain the sampled signal u(n), and calculate a root mean square voltage value for u(n) for each half period; S202: Transform the sequence U(t) to obtain its spectrum; perform a Fast Fourier Transform on the sequence U(t) to obtain the voltage fluctuation D(i), and calculate the instantaneous flicker value at that frequency: ; ; ; ; ; Where σ is the voltage fluctuation required to produce one unit of instantaneous flicker at this frequency. This value can be referenced from the IEC61000-4-15 standard, and is 0.407 under these conditions. S203: Calculate the instantaneous flicker value corresponding to the root mean square voltage sequence U(t); sum the instantaneous flicker values corresponding to each frequency, and obtain the instantaneous flicker value corresponding to U(t). ; S204: Calculate short-time flicker value; Short-time flicker value:
[0010]
[0011] ρ1 is 0.0314, ρ2 is 0.0525, ρ3 is 0.0657, ρ4 is 0.28, ρ5 is 0.08, Q i Q j Q k Q l Q m Q represents the instantaneous flicker values at 0.1%, 1%, 3%, 10%, and 50% of the time period, respectively. s If the limit is exceeded, mark the batch of integrated circuits and test them.
[0012] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. An artificial intelligence-based method for monitoring integrated circuit production, the method comprising the following steps: S1. Monitor the processing status of the production line, monitor the completion status of each part processing step, and correct any missing or incomplete steps; In step S1, to monitor whether any part processing steps are missing, a signaling method is proposed for monitoring: S101. In the initial processing stage, a Boolean signal is placed in this stage, and the Boolean signal contains only two data: 0 and 1. S102. In each processing stage, after a processing action occurs, a Boolean signal is incremented once, and a Boolean signal is sent to the next processing stage. The Boolean signal recorded in this stage is 0, and it increments to 1 after the processing action. S103. Before the next processing stage begins, detect the Boolean signal. If the Boolean signal is 1, it means that the integrated circuit was processed in the previous processing stage, so continue with the current processing stage. If the Boolean signal is 0, it means that there was a problem with the processing of the integrated circuit in the previous processing stage. In this case, transport the integrated circuit through another channel and transfer it back to the entry end of the previous processing stage. The reverse is also true for the Boolean signal, which only jumps between 0 and 1. Without changing the order of the processing stages, the Boolean signal received in each processing stage remains unchanged. Plot the Boolean signal sequence as a graph. S2. Monitor voltage fluctuations during production line processing and predict the production period based on real-time data; S3. Receive production information, schedule the work sequence of the production line within the factory, integrate the upstream and downstream of the production chain, and coordinate the scheduling between different manufacturers. S4. Monitor the inbound and outbound status after scheduling, and monitor transportation time and losses during transportation. 2.The integrated circuit production monitoring method based on artificial intelligence of claim 1, wherein: In step S2, the processing stage is electrically driven. The voltage monitoring stage is used to monitor the real-time voltage of the processing stage. In the power supply, the rated voltage is continuously input after being converted by a transformer. The input voltage is generally stable over a period of time. Based on a multi-functional instrument, it is detected whether the voltage meets the rated voltage requirements of the processing machinery, and alarms are triggered for voltage drop and overvoltage phenomena. The input voltage fluctuates at a certain moment. Now, the voltage flicker measurement method based on FFT is used to monitor the instantaneous voltage flicker value and record the batch number of the integrated circuit produced at the flicker moment for inspection to see if the instantaneous voltage flicker affects the processing. The instantaneous flicker value is monitored and calculated through the following steps. S201: Using the production time of each batch as the time period, calculate the root mean square voltage sequence U(t) within the period; sample the voltage signal u(t) at n points within each half period to obtain the sampled signal u(n), and calculate a root mean square voltage value for u(n) for each half period; S202: Transform the sequence U(t) to obtain the spectrum; perform a fast Fourier transform on the sequence U(t) to obtain the discrete spectrum, and divide twice the amplitude of the spectrum at the corresponding frequency by the amplitude of the decomposed DC component to obtain the voltage fluctuation D(i), and calculate the instantaneous flicker value at that frequency; S203: Calculate the instantaneous flicker value corresponding to the root mean square voltage sequence U(t); sum the instantaneous flicker values corresponding to each frequency, and obtain the instantaneous flicker value Q corresponding to U(t); S204: calculate short-time flicker value Q s After obtaining Q s , judge whether the instantaneous flicker value of the current processing link exceeds the limited area, which is obtained by monitoring the correlation between multiple production data and product quality; recheck the processing batch whose instantaneous flicker value exceeds the limited area to avoid integrated circuit quality problems; In the prediction unit, based on the production link production unit integrated circuit time α and production plan μ, the expected completion production time is calculated . 3.The integrated circuit production monitoring method based on artificial intelligence of claim 1, wherein: In step S3, based on real-time data and project schedule prediction results, the production process and transportation are further optimized; some production processes of integrated circuits have a sequential order, and some production processes can be scheduled within the scope, and the scheduling method is as follows; S301. Set the processing stage to W, the processing stage state to Wi, and the processing stage states to busy and idle respectively. The processing stage W is activated under the idle condition. S302. Retrieve integrated circuit batches that have completed the production process and are in the waiting stage; S303. Search all processing steps. Add processing steps with a fixed sequence to an ordered sequence of strongly constrained processing steps. Mark the processing steps of each batch of integrated circuits. Mark the completed processing steps and search for the processing steps that have not yet been performed. At this time, the processing steps before the idle processing steps are all busy or marked. Find the idle Wi and determine whether the processing needs to be transferred. If it needs to be transferred, determine the waiting time and the transfer processing time respectively. Based on the comparison results, submit a partial transfer application to the downstream factory to process and produce part of the batch of integrated circuits in advance. If there are no idle processing steps, jump to S304. S304. For processing steps with a fixed sequence, regardless of whether they are busy or idle, set them to busy and deactivate them. For interchangeable steps, set them to busy or idle according to their actual state and add them all to an ordered sequence of processing steps based on weak constraints. At the same time, mark the processing steps that have been completed for each batch of integrated circuits. Set the search to n steps forward each time. At this time, the processing steps before the idle processing steps are all busy or marked. Select the idle processing steps that do not need to be transferred for processing. If there are no idle processing steps, the search ends. The number of integrated circuits to be transferred to the processing steps is determined by the working efficiency of the work steps that are about to be transferred. Usually, it is set that the production of some transferred integrated circuits is completed before the next production cycle.
4. The integrated circuit production monitoring method based on artificial intelligence according to claim 1, characterized in that: In step S4, based on scheduling requirements, the inbound and outbound status is recorded, and the transportation process is monitored to predict arrival time and monitor losses during transportation. After the scheduling task is issued, the outbound unit receives the scheduling task order, breaks down the data to obtain the transfer destination and quantity, issues a transfer outbound task order, and records the outbound data. In the integrated circuit outbound transportation process, the transportation mode is monitored, and the transfer vehicle is accurately located based on the positioning system to estimate the transfer arrival time. During transportation, the loss unit promptly captures and uploads the loss information to the system so that it can be quickly dispatched from other sources to replenish the lost portion. When the next manufacturer conducts inbound inventory, the data from the outbound unit and the loss unit are integrated.
5. An artificial intelligence-based integrated circuit production monitoring system, applied to the artificial intelligence-based integrated circuit production monitoring method according to any one of claims 1-4, characterized in that: The production monitoring system includes a processing monitoring module, a scheduling module, and an inbound / outbound monitoring module. The processing monitoring module monitors each processing step according to the production process, obtains and corrects the real-time processing status of the production line, and predicts the processing time required for each processing step to complete its task based on the production line efficiency. The scheduling module receives the processing status of the production line and adjusts the processing procedures within the factory through a scheduling algorithm, integrating and connecting manufacturers across the entire production chain to optimize the processing sequence between different manufacturers. The inbound / outbound monitoring module monitors detailed information on finished products processed and shipped from the factory, and performs inbound inspection after the products are transferred to the next processing stage factory to monitor transportation losses.
6. The integrated circuit production monitoring system based on artificial intelligence according to claim 5, characterized in that: The processing monitoring module includes a process monitoring unit, a voltage monitoring unit, a correction unit, and a prediction unit; The process monitoring unit monitors the integrated circuit processing steps, acquires data from the processing machinery via sensors electrically connected to the machinery, and checks whether the integrated circuit has completed its processing action after each processing step. This unit monitors all integrated circuits to avoid omissions in processing steps. The voltage monitoring unit monitors the voltage status of the production line. The correction unit issues an abnormal alarm for abnormal temperatures on the production line and performs secondary processing to repair integrated circuits that have missed processing steps before they enter the next production step. The prediction unit estimates the completion time based on the current working efficiency of the production line and the production task of this batch of integrated circuits.
7. The integrated circuit production monitoring system based on artificial intelligence according to claim 5, characterized in that: The scheduling module includes an information receiving unit, a communication unit, a process adjustment unit, and a coordination unit; The information receiving unit receives information about processing progress from different manufacturers, providing basic data for system coordination; the communication unit connects upstream and downstream manufacturers, providing communication services and a platform for information exchange, reducing information barriers; the process adjustment unit adjusts the processing quantity and time based on the workload of each processing stage within the factory, arranging processing plans; the coordination unit optimizes the processing sequence based on the production time between manufacturers, reducing idle waiting time.
8. The integrated circuit production monitoring system based on artificial intelligence according to claim 5, characterized in that: The inbound and outbound monitoring module includes an outbound information unit, an inbound information unit, a transportation monitoring unit, and a loss unit; After receiving the scheduling task, the outbound information unit records the outbound information of integrated circuit parts. In the transportation monitoring unit, the location of the integrated circuit transportation and the estimated arrival time are located. The loss unit monitors the loss caused during packaging, transportation and handling, and submits a supplementary application for integrated circuit parts according to the expected delivery quantity. The inbound information unit records the received integrated circuit parts information and compares it with the outbound transportation data.
Citation Information
Patent Citations
Machining workshop scheduling and monitoring system and method
CN104657835A
System and method for real-time monitoring and analysis of power quality based on wavelet transform.
CN108872743A
Intelligent factory system based on Internet of Things technology and digital management technology
CN112286148A
