Carbon black production equipment health management method and system based on large model, equipment and medium

Abstract

The invention provides a carbon black production equipment health management method and system based on a large model, equipment and a medium, and belongs to the technical field of industrial Internet of Things, and the method comprises the steps: collecting equipment key parameters of carbon black production equipment, process parameters of a process pipeline and environmental parameters through an Internet of Things sensor, extracting features, and generating a structured feature vector; building and training a carbon black production equipment health management large model based on the TFT model; inputting the real-time structured feature vector into a carbon black production equipment health management large model, and calculating an equipment health assessment index based on output; synchronizing the extracted features to a twinborn model regularly through an edge computing node, and constructing a digital twinborn body; and generating a health state billboard based on the digital twinborn body, and visually displaying the equipment health assessment index, the output of the carbon black production equipment health management large model and the extracted features. According to the invention, intelligent monitoring and prediction of the carbon black equipment are realized, the health state is accurately evaluated, and the risk of unplanned shutdown is reduced.

Description

Technical Field

[0001] The present application belongs to the field of industrial Internet of Things technology, and specifically relates to a large-scale model-based health management method, system, equipment and medium for carbon black production equipment. Background Art

[0002] In the field of carbon black production, the production process is complex. The Industrial Internet of Things deploys sensor networks to monitor the operation of key equipment such as reactors, crushers, and coolers. However, traditional health management methods have the following shortcomings: First, the maintenance of traditional carbon black production equipment mostly relies on manual experience. Manual judgment of equipment failure is not only highly subjective and prone to errors, but also requires extremely high professional quality of personnel. For example, if the failure is judged only by the sound and temperature of the equipment operation, minor hidden dangers are easily overlooked, and the equipment is often damaged by the time it is discovered, affecting the production progress. Secondly, the existing equipment failure warning system is mostly based on simple threshold judgment and cannot capture the complex time-lag relationship between equipment status and process parameters. For example, if the temperature of the reactor is abnormal, the existing system only issues an overtemperature alarm and cannot predict the failure in advance based on the changes in process parameters, resulting in equipment damage and large downtime losses. Finally, the equipment operation data in carbon black production is complex and multi-source, and is currently only simply processed after collection. For example, the vibration signal is only used for vibration amplitude monitoring and is not sufficiently correlated with the process parameters, resulting in a waste of data value and an inability to provide strong support for equipment health assessment and fault diagnosis.

[0003] In summary, traditional carbon black production equipment management mainly relies on manual experience for equipment maintenance, lacks accurate monitoring of the equipment's real-time status and fault warning, resulting in inefficient equipment health management and difficulty meeting the needs of modern industrial intelligent production. Summary of the Invention

[0004] In a first aspect, an embodiment of the present application provides a large-scale model-based carbon black production equipment health management method, comprising the following steps: S1. Use IoT sensors to collect key parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters. Edge computing nodes then perform feature extraction on the collected data to generate structured feature vectors. S2. Build and train a large model for health management of carbon black production equipment based on the TFT model: Input is structured feature vector; The time-lag correlation between process parameters and key equipment parameter characteristics is captured through the lagged attention module; Output failure probability, process parameter deviation and health prediction score; S3. Input the real-time structured feature vector into the carbon black production equipment health management model and calculate the equipment health assessment indicators based on the output: failure probability, current health score, and health change rate; S4. Synchronize the extracted features to the twin model regularly through the edge computing node to build a digital twin; S5. Generate a health status dashboard based on the digital twin to visualize the equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features.

[0005] Furthermore, the specific steps of step S1 are as follows: S11. Deploy an infrared thermal imager on the outer wall of the reactor to collect the temperature field matrix ; Where H is the number of rows in the matrix and W is the number of columns in the matrix; S12. Embed vibration sensors in the bearing seats of the crusher and cooler to collect vibration acceleration signals ; Where N is the number of samples; S13. Deploy flow meters, oxygen analyzers, and temperature sensors in process pipelines to collect process parameter vectors. ; S14. Deploy dust sensors in the carbon black production workshop to collect environmental data: ; S15. Deploy edge computing nodes in the carbon black production workshop to perform feature extraction on the data collected by various IoT sensors: Calculate the time domain characteristics of the vibration acceleration signals of the crusher and cooler: mean and kurtosis ; After adding Hanning window to the vibration acceleration signals of the crusher and cooler, perform FFT calculation to calculate the frequency domain characteristics: FFT peak ; Define the superheated area for the temperature field of the reactor , calculate the proportion of overheating area and construct statistical characteristics of temperature field ; in, is the reference temperature, is the standard deviation of the temperature field; S16. Generate structured feature vector ; Where P is the process parameter, is the mean value of the vibration acceleration signal, represents the kurtosis of the vibration acceleration signal, represents the FFT peak value, represents the statistical characteristics of the temperature field, Represents environmental data.

[0006] Furthermore, the specific steps of step S2 are as follows: S21. Construct a large-scale model for health management of carbon black production equipment based on the TFT model; S22. Set structured feature vector is the input; S23. The delayed attention module is implemented by adding a time delay unit to the encoding layer of the TFT model and fusing the delayed features with the real-time features through gated integration. S24. Set the probability of classified failures and each type of failure as the first output, calculate the deviation of each parameter from the parameter standard value and use it as the second output, and use the health prediction score for the future set period as the third output; S25. Train the large model for health management of carbon black production equipment and send it to the edge computing node.

[0007] Furthermore, the specific steps of step S25 are as follows: S251. Collect structural feature vectors of carbon black production equipment under various operating conditions such as cold start, hot start, normal shutdown, and emergency shutdown, as well as various load changes to construct a data set; S252. Label the normal samples and fault samples in the data set, and set the fault samples to include the process parameters within the time [-t, t] before and after the process parameter changes; S253. Construct multi-task loss function;

[0008] in, 、 、 is the preset weight parameter, is the fault classification loss, is the process parameter deviation loss, It is the health score that predicts loss; S254. Use the dataset to train a large model for health management of carbon black production equipment, and use the Adam optimizer to optimize during the training process until the loss function converges or the maximum number of iterations is met.

[0009] Furthermore, the specific steps of step S3 are as follows: S31. Input the structured feature vector generated by the real-time data collection into the trained carbon black production equipment health management model to obtain the failure probability, process parameter deviation and health prediction score of each type of failure; S32. Filter out the fault type with the highest probability and calculate the current health score:

[0010] in, is the maximum failure probability, It is the comprehensive deviation of industrial parameters, which is calculated by the deviation of process parameters output by the model; 、 is the preset weight parameter; S33. Calculate the current health score based on the health prediction score:

[0011] Where k is the time window.

[0012] Furthermore, the specific steps of step S4 are as follows: S41. The edge computing node selects features whose mutual information is higher than a preset mutual information threshold from the extracted features;

[0013] in, Characterized by and health score The mutual information of is the preset mutual information threshold; S42. Compress the selected features according to a preset ratio and transmit them to the cloud using the MQTT protocol; S43. Building a digital twin in the cloud: The statistical characteristics of the temperature field Mapped as a surface thermal map of the reactor 3D model; Environmental data Convert to visual elements; S44. Current health score , Current Health Score The fault probability is bound to the twin metadata to complete the health indicator injection.

[0014] Furthermore, the specific steps of step S5 are as follows: S51. Build an equipment health status dashboard based on the digital twin, visualizing equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features. S52. On the equipment health status dashboard, the equipment status and environmental data represented by integrated process parameters and vibration acceleration signals are displayed simultaneously; S53. Map the current health score to the color of the device twin model to display the health status of the device, and overlay the fault type probability distribution on the surface of the device twin model.

[0015] In a second aspect, the present application also provides a large-scale model-based carbon black production equipment health management system, including: The data acquisition and feature extraction module is used to collect key equipment parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters through IoT sensors. It then performs feature extraction on the collected data through edge computing nodes to generate structured feature vectors. Model building and training module, used to build and train a large model for health management of carbon black production equipment based on the TFT model: Input is structured feature vector; The time-lag correlation between process parameters and key equipment parameter characteristics is captured through the lagged attention module; Output failure probability, process parameter deviation and health prediction score; The health assessment module is used to input real-time structured feature vectors into the carbon black production equipment health management model and calculate equipment health assessment indicators based on the output: failure probability, current health score and health change rate; The digital twin construction module is used to synchronize the extracted features to the twin model through edge computing nodes to build a digital twin; The visualization display module is used to generate a health status dashboard based on the digital twin, and to visualize the equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features.

[0016] In a third aspect, an embodiment of the present application further provides an electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the program, the steps of the large-model-based carbon black production equipment health management method as described in the first aspect are implemented.

[0017] In a fourth aspect, an embodiment of the present application further provides a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the large model-based carbon black production equipment health management method as described in the first aspect are implemented.

[0018] It can be seen from the above technical solutions that this application has the following advantages: The large-scale model-based health management method, system, equipment, and medium for carbon black production equipment provided in this application convert real-time data of carbon black production equipment into visualization indicators by integrating the Internet of Things, large models, and digital twins, thereby enabling rapid location of potential fault hazards, and presenting predictions and health trends in a dynamic visualization form to assist managers in anticipating risks; ultimately, the remaining life prediction of equipment and intelligent optimization of maintenance strategies are achieved, thereby reducing the risk of unplanned downtime and improving carbon black production efficiency and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for the description. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 This is a flow chart of the large-scale model-based carbon black production equipment health management method of the present invention.

[0021] Figure 2 Schematic diagram of the large-scale model-based health management system for carbon black production equipment of the present invention. DETAILED DESCRIPTION

[0022] The various embodiments of the present disclosure will be described in more detail below in the specific steps of the large-scale model-based carbon black production equipment health management method. The present disclosure can have various embodiments, and adjustments and changes can be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but rather that the present disclosure should be understood to cover all adjustments, equivalents, and / or alternatives that fall within the spirit and scope of the various embodiments of the present disclosure.

[0023] For example, in the field of carbon black production, although the Industrial Internet of Things has achieved a certain degree of operational monitoring of key equipment such as reactors, crushers, and coolers by deploying sensor networks, traditional health management methods for carbon black production equipment still have many drawbacks, mainly reflected in the following aspects: First, traditional equipment maintenance methods overly rely on manual experience. Manual diagnosis of equipment failures is highly subjective, prone to misjudgment, and requires extremely high professional expertise, resulting in high labor costs. For example, relying solely on operating equipment sound and temperature to identify a fault can easily overlook minor hidden dangers. By the time they are discovered, the equipment is often already damaged, impacting production progress. Second, current equipment failure warning systems mostly rely on simple threshold judgments, which fail to capture the complex time-lag relationships between equipment status and process parameters. For example, in the case of abnormal reactor temperature, existing systems only issue alarms when the temperature exceeds the specified limit and fail to predict failures based on changes in process parameters, leading to equipment damage and significant downtime costs. Third, carbon black production equipment generates a large amount of complex, multi-source data during operation. However, currently, this collected data undergoes only simple processing, failing to fully realize its value. For example, vibration signals are typically used only to monitor vibration amplitude without in-depth correlation with process parameters. This results in significant waste of data value and fails to provide effective support for equipment health assessment and fault diagnosis.

[0024] In summary, traditional carbon black production equipment management mainly relies on manual experience and lacks accurate monitoring of the equipment's real-time status and effective fault warning methods. This makes equipment health management inefficient and difficult to meet the requirements of modern industrial intelligent production, seriously affecting the production efficiency and economic benefits of carbon black production companies.

[0025] To address the above issues, this embodiment provides a large-scale model-based health management method for carbon black production equipment, which enables intelligent monitoring and prediction of carbon black equipment, accurately assesses its health status, and reduces the risk of unplanned downtime.

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0027] See also Figure 1 FIG. 1 is a flow chart of a method for health management of carbon black production equipment based on a large model in a specific embodiment, the method comprising the following steps: S1. Use IoT sensors to collect key parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters. Edge computing nodes then perform feature extraction on the collected data to generate structured feature vectors. It should be noted that the real-time monitoring of the equipment operating status and the preliminary processing of data were achieved, providing data support for subsequent model analysis; S2. Build and train a large model for health management of carbon black production equipment based on the TFT model: Input is structured feature vector; The time-lag correlation between process parameters and key equipment parameter characteristics is captured through the lagged attention module; Output failure probability, process parameter deviation and health prediction score; It should be noted that this enables the model to learn the complex relationship between equipment status and process parameters, making it possible to accurately predict faults and conduct health assessments; S3. Input the real-time structured feature vector into the carbon black production equipment health management model and calculate the equipment health assessment indicators based on the output: failure probability, current health score, and health change rate; It should be noted that real-time assessment of equipment health status is achieved, helping managers to understand equipment operating conditions in a timely manner; S4. Synchronize the extracted features to the twin model regularly through the edge computing node to build a digital twin; It should be noted that the digital twin, as a virtual mirror of the device, can reflect the actual operating status of the device in real time, providing an intuitive visualization tool for device management; S5. Generate a health status dashboard based on the digital twin, visualizing the equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features. It should be noted that the health status and related data of the equipment can be presented to the management personnel in an intuitive manner.

[0028] This embodiment integrates the Internet of Things, big models, and digital twins to achieve real-time monitoring of carbon black production equipment, fault warnings, and accurate health status assessments, thereby reducing the risk of unplanned downtime, improving production efficiency and equipment reliability, and reducing dependence on manual labor.

[0029] Furthermore, as a refinement and expansion of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process in this embodiment, another large model-based carbon black production equipment health management method is provided, which includes the following steps: S1. Use IoT sensors to collect key parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters. Edge computing nodes then perform feature extraction on the collected data to generate structured feature vectors. The specific steps of step S1 are as follows: S11. Deploy an infrared thermal imager on the outer wall of the reactor to collect the temperature field matrix ; Where H is the number of rows in the matrix and W is the number of columns in the matrix; S12. Embed vibration sensors in the bearing seats of the crusher and cooler to collect vibration acceleration signals ; Where N is the number of samples; S13. Deploy flow meters, oxygen analyzers, and temperature sensors in process pipelines to collect process parameter vectors. ; S14. Deploy dust sensors in the carbon black production workshop to collect environmental data: ; S15. Deploy edge computing nodes in the carbon black production workshop to perform feature extraction on the data collected by various IoT sensors: Calculate the time domain characteristics of the vibration acceleration signals of the crusher and cooler: mean and kurtosis ;

[0030]

[0031] wherein, represents the vibration acceleration signal, represents the sample number, represents the mean value of the vibration acceleration signal, represents the kurtosis of the vibration acceleration signal; The FFT calculation frequency domain features are performed on the vibration acceleration signals of the pulverizer and the cooler after the Hanning window is added: FFT peak ;

[0032] wherein, is the Hanning window, is the Fourier transform; The superheated region is defined for the temperature field of the reaction kettle , the superheated region proportion is calculated , and the temperature field statistical features are constructed ;

[0033] wherein, is the maximum function, is the standard deviation function; is the reference temperature, is the temperature field standard deviation; S16. Generating a structured feature vector ; wherein, P is the process parameter, is the mean value of the vibration acceleration signal, represents the kurtosis of the vibration acceleration signal, represents the FFT peak, represents the temperature field statistical features, represents the environmental data; S2. Constructing and training a carbon black production equipment health management large model based on the TFT model: Input with a structured feature vector; The time lag correlation between the process parameters and the equipment key parameter features is captured through the lag attention module; Output the failure probability, process parameter deviation degree and health prediction score; The specific steps of step S2 are as follows: S21. Constructing a carbon black production equipment health management large model based on the TFT model; S22. Set the structured feature vector as the input; S23. The lag attention module increases the time delay unit in the encoding layer of the TFT model, and realizes it by gate fusion of lag features and real-time features; Specifically as follows:

[0034] in, are process parameters used as query parameters, , It is the key parameter characteristic of the device as a key value parameter. ; is the attention head dimension, 、 is the learnable weight matrix, is a projectable learning matrix; is the query vector dimension, is the key-value vector dimension; is the lag mask matrix;

[0035] is the time lag parameter, is the optimal time lag parameter; Only allow attention Feature correlation within a time window; Optimal lag parameter Determine by maximizing the mutual information objective parameter:

[0036] in, is the mutual information function, which measures the statistical correlation between two variables; is the current lag process parameters; Indicates the mean value of the current vibration acceleration signal, Indicates the kurtosis of the current vibration acceleration signal, Indicates the current FFT peak value; Specifically, , Reflects the process conditions at a certain moment in the past and is used to predict the current equipment status; S24. Set the probability of classified failures and each type of failure as the first output, calculate the deviation of each parameter from the parameter standard value and use it as the second output, and use the health prediction score for the future set period as the third output; S25. Train the large model for health management of carbon black production equipment and send it to the edge computing node; The specific steps of step S25 are as follows: S251. Collect structural feature vectors of carbon black production equipment under various operating conditions such as cold start, hot start, normal shutdown, and emergency shutdown, as well as various load changes to construct a data set; S252. Label the normal samples and fault samples in the data set, and set the fault samples to include the process parameters within the time [-t, t] before and after the process parameter changes; S253. Construct multi-task loss function;

[0037] in, 、 、 is the preset weight parameter, is the fault classification loss, is the process parameter deviation loss, It is the health score that predicts loss; S254. Use the dataset to train a large model for health management of carbon black production equipment, and use the Adam optimizer to optimize during the training process until the loss function converges or the maximum number of iterations is met; S3. Input the real-time structured feature vector into the carbon black production equipment health management model and calculate the equipment health assessment indicators based on the output: failure probability, current health score, and health change rate; The specific steps of step S3 are as follows: S31. Input the structured feature vector generated by the real-time data collection into the trained carbon black production equipment health management model to obtain the failure probability, process parameter deviation and health prediction score of each type of failure; S32. Filter out the fault type with the highest probability and calculate the current health score:

[0038] in, is the maximum failure probability, It is the comprehensive deviation of industrial parameters, which is calculated by the deviation of process parameters output by the model; 、 is the preset weight parameter; For example, For health status, Triggering a yellow alert. An orange alert has been triggered. Triggering a red alert; S33. Calculate the current health score based on the health prediction score:

[0039] Where k is the time window; For example, k can be 12, if Trigger red alert at 1 minute / hour; S4. Synchronize the extracted features to the twin model regularly through the edge computing node to build a digital twin; The specific steps of step S4 are as follows: S41. The edge computing node selects features whose mutual information is higher than a preset mutual information threshold from the extracted features;

[0040] in, Characterized by and health score The mutual information of is the preset mutual information threshold; S42. Compress the selected features according to a preset ratio and transmit them to the cloud using the MQTT protocol; S43. Building a digital twin in the cloud: The statistical characteristics of the temperature field Mapped as a surface thermal map of the reactor 3D model; Environmental data Convert to visual elements; Specifically, the dust concentration is mapped to the particle flow density, and the accumulated operating time is mapped to the equipment wear texture; S44. Current health score , Current Health Score The fault probability is bound to the twin metadata to complete the health indicator injection; S5. Generate a health status dashboard based on the digital twin, visualizing the equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features; The specific steps of step S5 are as follows: S51. Build an equipment health status dashboard based on the digital twin, visualizing equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features. S52. On the equipment health status dashboard, the equipment status and environmental data represented by integrated process parameters and vibration acceleration signals are displayed simultaneously; S53. Map the current health score to the color of the device twin model to display the health status of the device, and overlay the fault type probability distribution on the surface of the device twin model.

[0041] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0042] like Figure 2As shown, the following is an embodiment of the large-scale model-based carbon black production equipment health management system provided by the embodiment of the present disclosure. This system and the large-scale model-based carbon black production equipment health management method of the above-mentioned embodiments belong to the same inventive concept. For details not fully described in the embodiment of the large-scale model-based carbon black production equipment health management system, please refer to the embodiment of the large-scale model-based carbon black production equipment health management method.

[0043] The system includes: The data acquisition and feature extraction module is used to collect key equipment parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters through IoT sensors. It then performs feature extraction on the collected data through edge computing nodes to generate structured feature vectors. Model building and training module, used to build and train a large model for health management of carbon black production equipment based on the TFT model: Input is structured feature vector; The time-lag correlation between process parameters and key equipment parameter characteristics is captured through the lagged attention module; Output failure probability, process parameter deviation and health prediction score; The health assessment module is used to input real-time structured feature vectors into the carbon black production equipment health management model and calculate equipment health assessment indicators based on the output: failure probability, current health score and health change rate; The digital twin construction module is used to synchronize the extracted features to the twin model through edge computing nodes to build a digital twin; The visualization display module is used to generate a health status dashboard based on the digital twin, and to visualize the equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features.

[0044] This embodiment effectively solves the problems of strong reliance on manual experience, delayed fault warning, and insufficient data utilization in the management of traditional carbon black production equipment through the interactive collaboration of the data acquisition and feature extraction module, the model construction and training module, the health assessment module, the digital twin construction module, and the visualization display module, thereby promoting the intelligent development of carbon black production.

[0045] The large-scale model-based carbon black production equipment health management method provided in the embodiment of the present application can be applied to electronic devices. Those skilled in the art will understand that the electronic device structure involved in the embodiment of the present invention does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently. In the embodiment of the present invention, the electronic device includes but is not limited to a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described and / or required herein.

[0046] The electronic device may include a processor, an external memory interface, an internal memory, a universal serial bus (USB) interface, a charging management module, a power management module, a battery, a wireless communication module, an audio module, a speaker, a microphone, a sensor module, a button, a camera, a display, and a SIM card interface, etc.

[0047] It is understood that the structures illustrated in the embodiments of the present application do not constitute specific limitations on the electronic device. In other embodiments of the present application, the electronic device may include more or fewer components than shown, or combine or separate certain components, or arrange the components differently. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0048] A processor may include one or more processing units, such as a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU). Different processing units may be independent devices or integrated into one or more processors.

[0049] The processor can be the nerve center and command center of the electronic device. The controller can generate operation control signals based on the instruction opcode and timing signal to complete the control of instruction fetching and execution.

[0050] The processor may also include a memory for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can store instructions or data that the processor has just used or is reusing. If the processor needs to use the instruction or data again, it can directly call it from the memory. This avoids repeated accesses, reduces processor latency, and thus improves system efficiency.

[0051] The above-mentioned electronic device realizes the large-scale model-based health management method of carbon black production equipment in this application, which collects the key equipment parameters of carbon black production equipment, process parameters of process pipelines and environmental parameters through Internet of Things sensors, and performs feature extraction on the collected data through edge computing nodes to generate structured feature vectors; constructs and trains a large model for health management of carbon black production equipment based on the TFT model: inputs a structured feature vector; captures the time-delayed correlation between process parameters and key equipment parameter characteristics through a delayed attention module; outputs the failure probability, process parameter deviation and health prediction score; and inputs the real-time structured feature vector into the health management of carbon black production equipment. The big model calculates the equipment health assessment indicators based on the output: failure probability, current health score and health change rate; the extracted features are synchronized to the twin model through the edge computing node at regular intervals to build a digital twin; based on the digital twin, a health status dashboard is generated to visualize the equipment health assessment indicators, the output of the carbon black production equipment health management big model and the extracted features. This technical solution achieves the beneficial effect of real-time monitoring of carbon black production equipment, fault warning and accurate health status assessment through the integration of the Internet of Things, big models and digital twins, reducing the risk of unplanned downtime, improving production efficiency and equipment reliability, and reducing human dependence.

[0052] The storage medium provided in this application stores a program product that can implement a large-scale model-based health management method for carbon black production equipment.

[0053] The health management method for carbon black production equipment based on a large model includes: collecting key equipment parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters through Internet of Things sensors, and performing feature extraction on the collected data through edge computing nodes to generate structured feature vectors; constructing and training a large model for health management of carbon black production equipment based on the TFT model: inputting a structured feature vector; capturing the time-lag correlation between process parameters and key equipment parameter characteristics through a delayed attention module; outputting failure probability, process parameter deviation, and health prediction score; inputting real-time structured feature vectors into the large model for health management of carbon black production equipment, and calculating equipment health assessment indicators based on the output: failure probability, current health score, and health change rate; regularly synchronizing the extracted features to the twin model through edge computing nodes to construct a digital twin; generating a health status dashboard based on the digital twin, and visually displaying equipment health assessment indicators, the output of the large model for health management of carbon black production equipment, and the extracted features.

[0054] In some possible embodiments, the large model-based carbon black production equipment health management method disclosed herein can be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps described in the above "Exemplary Method" section of this specification according to various exemplary embodiments of the present disclosure.

[0055] The storage medium of the present disclosure can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0056] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A carbon black production equipment health management method based on a large model, characterized in that: The steps include: S1. Use IoT sensors to collect key parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters. Edge computing nodes then perform feature extraction on the collected data to generate structured feature vectors. S2. Build and train a large model for health management of carbon black production equipment based on the TFT model: Input is structured feature vector; The time-lag correlation between process parameters and key equipment parameter characteristics is captured through the lagged attention module; Output failure probability, process parameter deviation and health prediction score; S3. Input the real-time structured feature vector into the carbon black production equipment health management model and calculate the equipment health assessment indicators based on the output: failure probability, current health score, and health change rate; S4. Synchronize the extracted features to the twin model regularly through the edge computing node to build a digital twin; S5. Generate a health status dashboard based on the digital twin to visualize the equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features.

2. The large model-based carbon black production equipment health management method according to claim 1, characterized in that: The specific steps of step S1 are as follows: S11. Deploy an infrared thermal imager on the outer wall of the reactor to collect the temperature field matrix ; Where H is the number of rows in the matrix and W is the number of columns in the matrix; S12. Embed vibration sensors in the bearing seats of the crusher and cooler to collect vibration acceleration signals ; Where N is the number of samples; S13. Deploy flow meters, oxygen analyzers, and temperature sensors in process pipelines to collect process parameter vectors. ; S14. Deploy dust sensors in the carbon black production workshop to collect environmental data: ； S15. Deploy edge computing nodes in the carbon black production workshop to perform feature extraction on the data collected by various IoT sensors: Calculate the time domain characteristics of the vibration acceleration signals of the crusher and cooler: mean and kurtosis ; After adding Hanning window to the vibration acceleration signals of the crusher and cooler, perform FFT calculation to calculate the frequency domain characteristics: FFT peak ; Define the superheated area for the temperature field of the reactor , calculate the proportion of overheating area and construct the statistical characteristics of temperature field ; in, is the reference temperature, is the standard deviation of the temperature field; S16. Generate structured feature vector ; Where P is the process parameter, is the mean value of the vibration acceleration signal, represents the kurtosis of the vibration acceleration signal, represents the FFT peak value, represents the statistical characteristics of the temperature field, Represents environmental data.

3. The large model-based carbon black production equipment health management method according to claim 2, characterized in that: The specific steps of step S2 are as follows: S21. Construct a large-scale model for health management of carbon black production equipment based on the TFT model; S22. Set structured feature vector is the input; S23. The delayed attention module is implemented by adding a time delay unit to the encoding layer of the TFT model and fusing the delayed features with the real-time features through gated integration. S24. Set the probability of classified failures and each type of failure as the first output, calculate the deviation of each parameter from the parameter standard value and use it as the second output, and use the health prediction score for the future set period as the third output; S25. Train the large model for health management of carbon black production equipment and send it to the edge computing node.

4. The large model-based carbon black production equipment health management method according to claim 3, characterized in that: The specific steps of step S25 are as follows: S251. Collect structural feature vectors of carbon black production equipment under various operating conditions such as cold start, hot start, normal shutdown, and emergency shutdown, as well as various load changes to construct a data set; S252. Label the normal samples and fault samples in the data set, and set the fault samples to include the process parameters within the time [-t, t] before and after the process parameter changes; S253. Construct multi-task loss function; in, 、 、 is the preset weight parameter, is the fault classification loss, is the process parameter deviation loss, It is the health score that predicts loss; S254. Use the dataset to train a large model for health management of carbon black production equipment, and use the Adam optimizer to optimize during the training process until the loss function converges or the maximum number of iterations is met.

5. The large model-based carbon black production equipment health management method according to claim 3, characterized in that: The specific steps of step S3 are as follows: S31. Input the structured feature vector generated by the real-time data collection into the trained carbon black production equipment health management model to obtain the failure probability, process parameter deviation and health prediction score of each type of failure; S32. Filter out the fault type with the highest probability and calculate the current health score: in, is the maximum failure probability, It is the comprehensive deviation of industrial parameters, which is calculated by the deviation of process parameters output by the model; 、 is the preset weight parameter; S33. Calculate the current health score based on the health prediction score: Where k is the time window.

6. The large model-based carbon black production equipment health management method according to claim 5, characterized in that: The specific steps of step S4 are as follows: S41. The edge computing node selects features whose mutual information is higher than a preset mutual information threshold from the extracted features; in, Features and health score The mutual information of is the preset mutual information threshold; S42. Compress the selected features according to a preset ratio and transmit them to the cloud using the MQTT protocol; S43. Building a digital twin in the cloud: The statistical characteristics of the temperature field Mapped as a surface thermal map of the reactor 3D model; Environmental data Convert to visual elements; S44. Current health score , Current Health Score The fault probability is bound to the twin metadata to complete the health indicator injection.

7. The large model-based carbon black production equipment health management method according to claim 6, characterized in that: The specific steps of step S5 are as follows: S51. Build an equipment health status dashboard based on the digital twin, visualizing equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features. S52. On the equipment health status dashboard, the equipment status and environmental data represented by integrated process parameters and vibration acceleration signals are displayed simultaneously; S53. Map the current health score to the color of the device twin model to display the health status of the device, and overlay the fault type probability distribution on the surface of the device twin model.

8. A carbon black production equipment health management system based on a large model, characterized in that: include: The data acquisition and feature extraction module is used to collect key equipment parameters of carbon black production equipment, process parameters of process pipelines, and environmental parameters through IoT sensors. It then performs feature extraction on the collected data through edge computing nodes to generate structured feature vectors. Model building and training module, used to build and train a large model for health management of carbon black production equipment based on the TFT model: Input is structured feature vector; The time-lag correlation between process parameters and key equipment parameter characteristics is captured through the lagged attention module; Output failure probability, process parameter deviation and health prediction score; The health assessment module is used to input real-time structured feature vectors into the carbon black production equipment health management model and calculate equipment health assessment indicators based on the output: failure probability, current health score and health change rate; The digital twin construction module is used to synchronize the extracted features to the twin model through edge computing nodes to build a digital twin; The visualization display module is used to generate a health status dashboard based on the digital twin, and to visualize the equipment health assessment indicators, the output of the carbon black production equipment health management model, and the extracted features.

9. An electronic device, characterized in that: It includes a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the large model-based health management method for carbon black production equipment as described in any one of claims 1 to 7 are implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the large model-based carbon black production equipment health management method as described in any one of claims 1 to 7 are implemented.

Images