Industrial scene-oriented data acquisition system and acquisition method thereof

Abstract

The invention discloses an industrial scene-oriented data acquisition system and an industrial scene-oriented data acquisition method. According to the invention, through the multi-level dynamic optimization design, the data acquisition efficiency and the system reliability in a complex environment are significantly improved. The system dynamically allocates thread resources based on the real-time communication state of the equipment, realizes stable throughput under a high-concurrency scene in combination with core binding and polling scheduling strategies, and ensures that millisecond-level response is still maintained when 10000-level equipment is accessed. The memory preloading and protocol template technology eliminates jitter during operation, the hierarchical resource isolation mechanism provides deterministic guarantee for key control instructions, and interruption of the production process due to data delay or resource competition is avoided. The data flow delay is further reduced through multi-protocol efficient analysis and intelligent cache management, seamless compatibility of heterogeneous equipment in a hybrid networking scene is supported, and the strict requirements of continuous production industries such as steel and chemical engineering for real-time performance and stability are met.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial data acquisition, and specifically relates to a data acquisition system and acquisition method for industrial scenarios. Background Art

[0002] An industrial data acquisition system is an integrated system used to monitor and control various parameters in industrial production processes in real time. It utilizes hardware devices such as sensors, actuators, and data acquisition cards, along with corresponding software platforms, to automatically collect, transmit, process, and store production site data. The system collects key process parameters such as temperature, pressure, humidity, flow rate, and rotational speed, providing accurate data support for production management and optimization. The application of industrial data acquisition systems helps improve production efficiency, reduce energy consumption, and ensure safe equipment operation. It also lays the foundation for enterprises' intelligent transformation and the development of the Industrial Internet, making it a crucial component of modern industrial automation and informatization.

[0003] However, existing industrial data collection technologies often suffer from rigid resource allocation in high-concurrency scenarios due to their fixed thread pool design, which can easily lead to thread starvation or resource idling when device communication delays fluctuate. Static protocol parsing and caching strategies are difficult to adapt to multi-source heterogeneous devices, resulting in excessive real-time instruction processing delays. There is a lack of global health assessment and dynamic tuning mechanisms, and the technology relies on manual experience to deal with device anomalies, resulting in delayed fault responses and rising operation and maintenance costs. Summary of the Invention

[0004] The purpose of the present invention is to provide a data acquisition system and method for industrial scenarios in order to solve the above-mentioned problems.

[0005] The technical solution adopted by the present invention is as follows: a data acquisition system and acquisition method for industrial scenarios, the system comprising: a data acquisition module, a data processing module, a data transmission module and an intelligent analysis and self-optimization module;

[0006] The intelligent analysis and self-optimization module is internally provided with a real-time monitoring and diagnosis engine, a predictive maintenance model module, a dynamic parameter optimizer and a self-healing actuator;

[0007] The data acquisition module is internally provided with a multi-threaded scheduling engine and a resource allocation controller. The data transmission module is internally provided with a protocol conversion engine and a hash matching optimizer.

[0008] The output end of the data acquisition module is connected to the input end of the data processing module through a high-speed data channel.

[0009] The output end of the data processing module is connected to the input end of the data transmission module, and the classified and marked data is pushed to the protocol conversion engine for standardized packaging;

[0010] The output end of the data transmission module is connected to an external system via an industrial bus or a network protocol interface.

[0011] The input end of the intelligent analysis and self-optimization module is respectively connected to the thread pool status monitoring end of the data acquisition module, the cache hit rate feedback end of the data processing module and the protocol conflict rate statistics end of the data transmission module, so as to collect the operating indicators of each module in real time; its output end is reversely connected to the thread pool parameter configuration end of the data acquisition module, the weight coefficient adjustment end of the data processing module and the hash table reconstruction end of the data transmission module through the control interface.

[0012] In a preferred embodiment, the multi-threaded scheduling engine dynamically builds an elastic thread pool based on the device communication delay, and uses an adaptive algorithm to adjust the number of threads by collecting device response time indicators in real time. For high-latency devices, the occupied threads are reduced proportionally and resources are recovered; for low-latency devices, the number of threads is expanded to improve throughput. The engine binds threads to specified physical cores through system-level APIs to eliminate the performance loss caused by cross-core cache synchronization, and uses a weighted polling strategy to allocate tasks to ensure load balancing of multi-device communications. When a device response timeout or resource competition is detected, the thread migration mechanism is automatically triggered to reallocate tasks to low-load cores to maintain millisecond-level real-time performance. The thread pool capacity is dynamically set according to the device type and communication protocol. Industrial bus devices are allocated intensive computing threads, and sensor devices enable a lightweight thread model.

[0013] In a preferred embodiment, the resource allocation controller preloads protocol parsing templates, cache queues, and device topology data during the system startup phase, and uses memory locking technology to prevent critical data from being swapped out of physical memory. A hierarchical resource pool design is adopted to divide independent memory areas and dedicated CPU cores for real-time control instructions, ensuring zero preemption delay for high-priority tasks. The controller monitors the memory paging frequency and CPU cache hit rate in real time, and dynamically adjusts the amount of preloaded data and core binding strategy. For sudden load scenarios, the emergency resource channel is enabled, low-priority task resources are temporarily borrowed, and a rollback timer is set. After the load drops, the initial allocation state is automatically restored, achieving dual guarantees of resource utilization and stability.

[0014] In a preferred embodiment, the real-time monitoring and diagnosis engine collects multi-dimensional indicators such as device communication delay, thread pool load, cache hit rate, and protocol parsing efficiency in real time through a streaming data processing framework, and combines time series database storage with sliding window calculation to generate a system performance trend map. Anomaly detection is based on historical baseline modeling and real-time fluctuation analysis to identify abnormal signals such as sudden increases in communication delay and protocol parsing errors, and traces the root cause of the problem through multi-dimensional data correlation. When resource allocation is abnormal or device communication is interrupted, the engine triggers cross-module collaboration instructions to ensure that problem location and optimization decision-making are carried out simultaneously.

[0015] In a preferred embodiment, the predictive maintenance model module uses dynamic weights to fuse real-time data and historical operating characteristics of the equipment to quantitatively assess the health status of the equipment. The health index is calculated by integrating sensor indicators, environmental parameters, and system interaction logs, and the weight distribution is dynamically adjusted according to the equipment type, operating load, and failure risk. When the health index is lower than the preset threshold, the system automatically triggers maintenance warnings in a hierarchical manner, responding step by step from recommended inspections to forced shutdowns. At the same time, the module combines the equipment degradation model to predict the remaining life of key components, and links the resource management system to optimize spare parts scheduling, reducing the impact of unplanned shutdowns on the production process;

[0016] The predictive maintenance model module dynamically combines multi-dimensional real-time data with equipment characteristics to achieve quantitative assessment and adaptive adjustment of health status. The formula is defined as follows:

[0017]

[0018] Where CurrentMetric_i is the real-time measurement value of the metric of the i-th type of equipment (such as temperature, vibration amplitude, current fluctuation, etc.);

[0019] BaselineMetric_i is the baseline value of the i-th indicator in the healthy state of the equipment (the initial value is determined by the equipment factory parameters or historical golden period data); i is the dynamic weight coefficient of the i-th category indicator, satisfying ∑wi=1. -λΔt is the time decay factor, where λ is the decay rate parameter, and Δt is the interval between the current time and the start time of the indicator anomaly.

[0020] In a preferred embodiment, the dynamic parameter optimizer autonomously adjusts system operating parameters based on real-time data analysis and machine learning models. A reinforcement learning framework is used to construct a state-action mapping relationship. Throughput, latency, and resource utilization are optimized, dynamically adjusting thread pool capacity, cache policy priority, or protocol parser resource allocation. Historical parameter combinations are associated with performance indicators and stored, automatically reverting to a stable configuration when new parameters cause performance fluctuations. This optimization process balances short-term efficiency gains with long-term system stability, preventing local optimal solutions from leading to global resource imbalances.

[0021] In a preferred embodiment, the self-healing actuator converts diagnostic and optimization results into control instructions, executing multi-level repair actions. Soft failures, such as protocol parsing anomalies, automatically load a backup template or switch to a redundant communication link. Hardware failures send a reset command via the industrial bus or trigger a physical protection mechanism. Compute nodes are dynamically scaled up or down at the resource level, flexibly allocating computing power based on load forecasts. Sensitive operations are simulated and verified using a digital twin system before execution to ensure safety and reliability, preventing secondary failures caused by misoperation.

[0022] In a preferred embodiment, the protocol conversion engine precompiles the fixed fields of industrial protocols such as Modbus and OPCUA into binary structures, completes syntax tree construction and machine code generation in the initialization phase, and improves runtime parsing efficiency to the level of direct memory access. It supports hot-plug extension of protocols, and new protocols define field offsets, verification rules and conversion logic through description files. The engine automatically generates an adapter parser and injects it into the runtime environment. For non-standard protocols, dynamic parsing mode is enabled, and the pre-stored template library is matched based on the first packet feature code. When there is no hit, the deep learning model is started to infer the protocol structure, and the conversion results are fed back to the template library to realize self-learning iteration. Binary template matching adopts a field weight accumulation algorithm, and key control fields are given higher weight values to ensure the accuracy of protocol recognition.

[0023] In a preferred embodiment, the hash matching optimizer constructs a direct mapping from the protocol signature to the parsing function, extracts the function code, address field and control symbol combination of the protocol header to generate a 128-bit hash key. A multi-layer hash table structure is adopted. The first layer diverts traffic in a coarse-grained manner according to the protocol type, and the second layer resolves conflicts through linear detection. The maximum detection depth is set to 15% of the hash table capacity. The optimizer counts the hash collision rate and matching delay of each protocol traffic in real time. When the collision rate exceeds the threshold, it automatically reconstructs the hash function parameters or expands the table capacity. A bypass detection channel is designed for high-concurrency scenarios. When the hash match of the first packet fails, it is transferred to the co-processing unit for deep analysis, and the results are asynchronously backfilled into the main hash table to ensure that sub-millisecond matching efficiency is always maintained during traffic peaks.

[0024] In a preferred embodiment, the data processing module establishes a three-level data management system, implements a preemptive transmission channel for real-time control instructions, and ensures that the entire link from caching to forwarding is completed within 5ms. A double-buffered queue design is used for key process data. When the main queue is full, it automatically switches to the backup queue and triggers asynchronous persistence operations at the same time. An improved LRU-K algorithm is introduced for historical data storage, which calculates the cache value based on access frequency, time locality and device weight, and dynamically eliminates low-value data. The weight calculation engine integrates the device criticality score, data timeliness attenuation factor and business strategy coefficient, and balances real-time response and storage cost through adjustable parameters. The real-time coefficient in the formula decays exponentially according to the difference in data generation timestamps, and the importance coefficient is dynamically generated through equipment operation and maintenance records and production process dependencies.

[0025] in:

[0026] -Level 1 (highest level): Real-time control instructions (such as emergency stop and parameter adjustment) must be cached and triggered within <5ms.

[0027] -Level 2 (Intermediate): Critical process data (such as temperature and pressure sensor values), with an allowed delay of ≤100ms.

[0028] -Level 3 (low level): historical records or redundant data, using the LRU elimination strategy to free up space.

[0029] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0030] 1. In the present invention, the data collection efficiency and system reliability in complex environments are significantly improved through multi-level dynamic optimization design. The system dynamically allocates thread resources based on the real-time communication status of the equipment, and combines core binding and polling scheduling strategies to achieve stable throughput in high-concurrency scenarios, ensuring millisecond-level response when tens of thousands of devices are connected. Pre-loaded memory and protocol template technology eliminate runtime jitter, and the hierarchical resource isolation mechanism provides deterministic protection for key control instructions, avoiding interruptions in the production process due to data delays or resource competition. Efficient multi-protocol parsing and intelligent cache management further reduce data flow delays, support seamless compatibility in mixed networking scenarios of heterogeneous equipment, and meet the stringent real-time and stability requirements of continuous production industries such as steel and chemical industries.

[0031] 2. The present invention breaks through the passive response mode of traditional data collection gateways and realizes closed-loop optimization of the entire link from data collection to autonomous decision-making. The built-in health assessment model predicts equipment failures through multi-dimensional data analysis and triggers maintenance instructions in advance to reduce unplanned downtime losses. The self-optimization module adjusts key parameters such as thread pools and cache strategies in real time, and combines the self-healing mechanism to quickly repair communication anomalies or protocol errors, reducing the frequency of manual intervention and operation and maintenance costs. This proactive operation and maintenance capability enables the system to adapt to dynamic industrial environments such as equipment aging and load fluctuations, extending the equipment life cycle while ensuring production efficiency, and providing a highly reliable data base for intelligent manufacturing upgrades. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a block diagram of the overall system of the present invention;

[0033] Figure 2 This is a system block diagram of the intelligent analysis and self-optimization module in the present invention;

[0034] Figure 3 This is a system block diagram of the data acquisition module in the present invention;

[0035] Figure 4 This is a system block diagram of the data transmission module in the present invention. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0037] Reference Figure 1-4 ,

[0038] A data acquisition system and method for industrial scenarios, the system includes: a data acquisition module, a data processing module, a data transmission module and an intelligent analysis and self-optimization module;

[0039] The intelligent analysis and self-optimization module is internally configured with a real-time monitoring and diagnosis engine, a predictive maintenance model module, a dynamic parameter optimizer, and a self-healing actuator;

[0040] The data acquisition module is internally configured with a multi-threaded scheduling engine and a resource allocation controller.

[0041] The data transmission module is internally provided with a protocol conversion engine and a hash matching optimizer.

[0042] The multi-threaded scheduling engine dynamically builds an elastic thread pool based on device communication latency, collects device response time indicators in real time, and uses an adaptive algorithm to adjust the number of threads. For high-latency devices, the occupied threads are reduced proportionally and resources are reclaimed; for low-latency devices, the number of threads is expanded to improve throughput. The engine binds threads to specified physical cores through system-level APIs to eliminate the performance loss caused by cross-core cache synchronization, and uses a weighted polling strategy to allocate tasks to ensure load balancing of multi-device communications. When a device response timeout or resource competition is detected, the thread migration mechanism is automatically triggered to reallocate tasks to low-load cores to maintain millisecond-level real-time performance. The thread pool capacity is dynamically set according to the device type and communication protocol. Industrial bus devices are allocated intensive computing threads, while sensor devices enable a lightweight thread model.

[0043] The resource allocation controller preloads protocol parsing templates, cache queues, and device topology data during system startup, using memory locking technology to prevent critical data from being swapped out of physical memory. A hierarchical resource pool design allocates independent memory areas and dedicated CPU cores for real-time control instructions, ensuring zero preemption delay for high-priority tasks. The controller monitors memory paging frequency and CPU cache hit rate in real time, dynamically adjusting the amount of preloaded data and core binding strategies. For sudden load scenarios, an emergency resource channel is enabled to temporarily borrow low-priority task resources and set a rollback timer. Once the load subsides, the initial allocation state is automatically restored, ensuring both resource utilization and stability.

[0044] The real-time monitoring and diagnosis engine uses a streaming data processing framework to collect multi-dimensional metrics such as device communication latency, thread pool load, cache hit rate, and protocol parsing efficiency. Combined with time-series database storage and sliding window calculations, it generates system performance trend maps. Anomaly detection, based on historical baseline modeling and real-time fluctuation analysis, identifies abnormal signals such as sudden increases in communication latency and protocol parsing errors, and traces the root cause of the problem through multi-dimensional data correlation. When resource allocation anomalies occur or device communication is interrupted, the engine triggers cross-module coordination commands to ensure simultaneous problem identification and optimization decisions.

[0045] The predictive maintenance model module uses dynamic weighting to integrate real-time equipment data with historical operating characteristics to quantitatively assess the health status of equipment. The health index calculation integrates sensor indicators, environmental parameters, and system interaction logs, and the weight distribution is dynamically adjusted based on equipment type, operating load, and failure risk. When the health index falls below the preset threshold, the system automatically triggers maintenance warnings in a graded manner, with a step-by-step response from recommended inspections to forced shutdowns. At the same time, the module combines equipment degradation models to predict the remaining life of key components, and links with the resource management system to optimize spare parts scheduling, reducing the impact of unplanned downtime on production processes.

[0046] The priority calculation formula is:

[0047] Weight = real-time coefficient + (1-alpha) times importance coefficient

[0048] Among them, `α∈[0,1]` is an adjustable parameter (default value is 0.7), the real-time coefficient is calculated based on the difference between the data timestamp and the system clock (such as the exponential decay model), and the importance coefficient is configured by the user or automatically evaluated by machine learning to assess the criticality of the device.

[0049] The predictive maintenance model module dynamically combines multi-dimensional real-time data with equipment characteristics to achieve quantitative assessment and adaptive adjustment of health status. The formula is defined as follows:

[0050]

[0051] Where CurrentMetric_i is the real-time measurement value of the metric of the i-th type of equipment (such as temperature, vibration amplitude, current fluctuation, etc.);

[0052] BaselineMetric_i is the baseline value of the i-th indicator in the healthy state of the equipment (the initial value is determined by the equipment's factory parameters or historical golden period data);

[0053] w i is the dynamic weight coefficient of the i-th category indicator, satisfying ∑wi=1; e -λΔt is the time decay factor, where λ is the decay rate parameter, and Δt is the interval between the current time and the start time of the indicator anomaly;

[0054] The dynamic parameter optimizer autonomously adjusts system operating parameters based on real-time data analysis and machine learning models. Using a reinforcement learning framework to construct state-action mappings, it dynamically adjusts thread pool capacity, cache policy priorities, and protocol parser resource allocation, optimizing throughput, latency, and resource utilization. Historical parameter combinations are associated and stored with performance metrics, automatically reverting to a stable configuration when new parameters cause performance fluctuations. The optimization process balances short-term efficiency gains with long-term system stability, preventing local optimal solutions from leading to global resource imbalances.

[0055] The self-healing actuator converts diagnostic and optimization results into control instructions, executing multi-level repair actions. Soft failures, such as protocol parsing anomalies, automatically load a backup template or switch to a redundant communication link. Hardware failures send reset commands via the industrial bus or trigger physical protection mechanisms. Compute nodes are dynamically scaled up and down at the resource level, elastically allocating computing power based on load forecasts. Sensitive operations are simulated and verified using a digital twin system before execution to ensure safety and reliability, preventing secondary failures caused by misoperation.

[0056] The protocol conversion engine pre-compiles the fixed fields of industrial protocols such as Modbus and OPCUA into binary structures, completes syntax tree construction and machine code generation during the initialization phase, and improves runtime parsing efficiency to the level of direct memory access. It supports hot-swappable protocol expansion, and new protocols define field offsets, verification rules, and conversion logic through description files. The engine automatically generates an adapter parser and injects it into the runtime environment. For non-standard protocols, dynamic parsing mode is enabled, and the pre-stored template library is matched based on the first packet feature code. If a hit is not found, the deep learning model is started to infer the protocol structure, and the conversion results are fed back to the template library to achieve self-learning iteration. Binary template matching uses a field weight accumulation algorithm, and key control fields are given higher weight values to ensure the accuracy of protocol recognition.

[0057] The hash matching optimizer constructs a direct mapping from the protocol signature to the parsing function, extracts the function code, address field, and control symbol combination of the protocol header to generate a 128-bit hash key. A multi-layer hash table structure is adopted. The first layer diverts traffic coarsely by protocol type, and the second layer resolves conflicts through linear detection. The maximum detection depth is set to 15% of the hash table capacity. The optimizer counts the hash collision rate and matching delay of each protocol traffic in real time. When the collision rate exceeds the threshold, it automatically reconstructs the hash function parameters or expands the table capacity. A bypass detection channel is designed for high-concurrency scenarios. When the hash match of the first packet fails, it is transferred to the co-processing unit for in-depth analysis. The results are asynchronously backfilled into the main hash table to ensure that sub-millisecond matching efficiency is always maintained during traffic peaks.

[0058] The data processing module establishes a three-level data management system and implements a preemptive transmission channel for real-time control instructions to ensure that the entire link from caching to forwarding is completed within 5ms. A double-buffered queue design is used for key process data. When the main queue is full, it automatically switches to the backup queue and triggers asynchronous persistence operations at the same time. An improved LRU-K algorithm is introduced for historical data storage. It calculates the cache value based on access frequency, time locality and device weight, and dynamically eliminates low-value data. The weight calculation engine integrates the device criticality score, data timeliness attenuation factor and business strategy coefficient, and balances real-time response and storage cost through adjustable parameters. The real-time coefficient in the formula decays exponentially according to the difference in data generation timestamps, and the importance coefficient is dynamically generated through equipment operation and maintenance records and production process dependencies.

[0059] in:

[0060] -Level 1 (highest level): Real-time control instructions (such as emergency stop and parameter adjustment) must be cached and triggered within <5ms.

[0061] -Level 2 (Intermediate): Critical process data (such as temperature and pressure sensor values), with an allowed delay of ≤100ms.

[0062] -Level 3 (low level): historical records or redundant data, using the LRU elimination strategy to free up space.

[0063] From the above we can know:

[0064] In the present invention, through multi-level dynamic optimization design, the data collection efficiency and system reliability in complex environments are significantly improved. The system dynamically allocates thread resources based on the real-time communication status of the equipment, and combines core binding and polling scheduling strategies to achieve stable throughput in high-concurrency scenarios, ensuring that millisecond-level response is maintained when tens of thousands of devices are connected. Pre-loaded memory and protocol template technology eliminate runtime jitter, and the hierarchical resource isolation mechanism provides deterministic protection for key control instructions, avoiding production process interruptions due to data delays or resource competition. Efficient multi-protocol parsing and intelligent cache management further reduce data flow delays, support seamless compatibility in mixed networking scenarios of heterogeneous equipment, and meet the stringent real-time and stability requirements of continuous production industries such as steel and chemical industries.

[0065] In this invention, the passive response mode of the traditional data collection gateway is broken through to achieve full-link closed-loop optimization from data collection to autonomous decision-making. The built-in health assessment model predicts equipment failures through multi-dimensional data analysis, triggers maintenance instructions in advance to reduce unplanned downtime losses. The self-optimization module adjusts key parameters such as thread pools and cache strategies in real time, and combines the self-healing mechanism to quickly repair communication anomalies or protocol errors, reducing the frequency of manual intervention and operation and maintenance costs. This proactive operation and maintenance capability enables the system to adapt to dynamic industrial environments such as equipment aging and load fluctuations, extending the equipment life cycle while ensuring production efficiency, and providing a highly reliable data base for intelligent manufacturing upgrades.

[0066] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0067] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A data acquisition system and method for industrial scenarios, characterized by: The system includes: a data acquisition module, a data processing module, a data transmission module and an intelligent analysis and self-optimization module; The intelligent analysis and self-optimization module is internally provided with a real-time monitoring and diagnosis engine, a predictive maintenance model module, a dynamic parameter optimizer and a self-healing actuator; The data acquisition module is internally provided with a multi-threaded scheduling engine and a resource allocation controller. The data transmission module is internally provided with a protocol conversion engine and a hash matching optimizer; The output end of the data acquisition module is connected to the input end of the data processing module through a high-speed data channel. The output end of the data processing module is connected to the input end of the data transmission module, and the classified and marked data is pushed to the protocol conversion engine for standardized packaging; The output end of the data transmission module is connected to an external system via an industrial bus or a network protocol interface; The input end of the intelligent analysis and self-optimization module is respectively connected to the thread pool status monitoring end of the data acquisition module, the cache hit rate feedback end of the data processing module and the protocol conflict rate statistics end of the data transmission module, so as to collect the operating indicators of each module in real time; its output end is reversely connected to the thread pool parameter configuration end of the data acquisition module, the weight coefficient adjustment end of the data processing module and the hash table reconstruction end of the data transmission module through the control interface.

2. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The multi-threaded scheduling engine dynamically builds an elastic thread pool based on device communication latency, collects device response time metrics in real time, and uses an adaptive algorithm to adjust the number of threads. For high-latency devices, it proportionally reduces the number of threads occupied and reclaims resources; for low-latency devices, it expands the number of threads to improve throughput. The engine binds threads to specified physical cores through system-level APIs, eliminating the performance loss caused by cross-core cache synchronization. At the same time, it adopts a weighted polling strategy to allocate tasks to ensure load balancing of multi-device communications. When a device response timeout or resource competition is detected, the thread migration mechanism is automatically triggered to reallocate tasks to low-load cores.

3. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The resource allocation controller preloads the protocol parsing template, cache queue and device topology data during the system startup phase, and prevents key data from being swapped out of the physical memory through memory locking technology.

4. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The real-time monitoring and diagnosis engine collects multi-dimensional indicators such as device communication delay, thread pool load, cache hit rate and protocol parsing efficiency in real time through a streaming data processing framework, and generates a system performance trend map by combining time series database storage and sliding window calculation; anomaly detection is based on historical baseline modeling and real-time fluctuation analysis to identify abnormal signals such as sudden increases in communication delay and protocol parsing errors, and trace the root cause of the problem through multi-dimensional data correlation.

5. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The predictive maintenance model module uses dynamic weighting to integrate real-time equipment data with historical operating characteristics to quantitatively assess equipment health. The health index is calculated based on comprehensive sensor indicators, environmental parameters, and system interaction logs, and the weight distribution is dynamically adjusted based on equipment type, operating load, and failure risk. When the health index falls below a preset threshold, the system automatically triggers maintenance warnings in a hierarchical manner, with responses ranging from recommended inspections to forced shutdowns. The predictive maintenance model module dynamically combines multi-dimensional real-time data with equipment characteristics to achieve quantitative assessment and adaptive adjustment of health status; the formula is defined as follows: Where CurrentMetric_i is the real-time measurement value of the metric of the i-th type of equipment; BaselineMetric_i is the baseline value of the i-th indicator in the health status of the device; w i is the dynamic weight coefficient of the i-th category indicator, satisfying ∑wi=1; e -λΔt is the time decay factor, where λ is the decay rate parameter, and Δt is the interval between the current time and the start time of the indicator anomaly.

6. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The dynamic parameter optimizer autonomously adjusts system operating parameters based on real-time data analysis and machine learning models; constructs a state-action mapping relationship through a reinforcement learning framework, dynamically adjusts thread pool capacity, cache strategy priority, or protocol parsing resource allocation with throughput, latency, and resource utilization as optimization targets; historical parameter combinations are associated with performance indicators and stored, automatically falling back to a stable configuration when new parameters cause performance fluctuations; the optimization process takes into account both short-term efficiency improvements and long-term system stability, avoiding global resource imbalances caused by local optimal solutions.

7. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The self-healing actuator converts the diagnosis and optimization results into control instructions and executes multi-level repair actions; when the soft fault protocol parsing is abnormal, it automatically loads the backup template or switches the redundant communication link; In the event of a hardware failure, a reset command is sent through the industrial bus or a physical protection mechanism is triggered; at the resource level, computing nodes are dynamically scaled up or down, and computing power is flexibly allocated based on load prediction results.

8. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The protocol conversion engine precompiles the fixed fields of Modbus and OPCUA industrial protocols into binary structures, completes syntax tree construction and machine code generation in the initialization phase, and improves runtime parsing efficiency to the level of direct memory access; it supports hot-swappable protocol extensions, and new protocols define field offsets, verification rules, and conversion logic through description files. The engine automatically generates an adapter parser and injects it into the runtime environment.

9. The data acquisition system and method for industrial scenarios according to claim 1, characterized in that: The hash matching optimizer constructs a direct mapping from the protocol signature to the parsing function, extracts the function code, address field and control symbol combination of the protocol header to generate a 128-bit hash key; A multi-layer hash table structure is used. The first layer performs coarse-grained traffic diversion based on protocol type, and the second layer resolves conflicts through linear probing, with a maximum probing depth set at 15% of the hash table capacity. The optimizer calculates the hash collision rate and matching delay of each protocol traffic in real time. When the collision rate exceeds a threshold, it automatically reconstructs the hash function parameters or expands the table capacity. The data processing module establishes a three-level data management system, implements a preemptive transmission channel for real-time control instructions, and ensures that the entire cache-to-forwarding link is completed within 5ms. Key process data adopts a double-buffer queue design. When the main queue is full, it automatically switches to the backup queue and triggers asynchronous persistence operations at the same time. The historical data storage introduces an improved LRU-K algorithm, which calculates the cache value based on access frequency, time locality and device weight, and dynamically eliminates low-value data. in: -Level 1, the highest level: real-time control instructions, which must be cached and triggered within <5ms; -Level 2 Intermediate: critical process data, with an allowable delay of ≤100ms; -Level 3 low level: historical records or redundant data, using the LRU elimination strategy to free up space.

10. A data collection method for industrial scenarios, characterized by: The method runs the industrial scenario-oriented data acquisition system according to any one of claims 1 to 9.

Images