A production line informatization control method and system based on the Internet of Things

Abstract

The application discloses a production line informationization control method and system based on the Internet of Things, which comprises the following steps: acquiring equipment state monitoring information, raw material inventory records and production progress tracking logs, uniformly formatting and merging multi-source data to form a basic data group, completing data fusion through sharding storage and parallel processing to obtain a multi-source data set, extracting a key data subset based on update frequency and transmission time delay sorting, analyzing bottleneck links in combination with the production progress tracking logs to generate resource demand feature records, performing dynamic resource allocation analysis based on the feature and the state monitoring information to obtain a resource allocation plan, performing inventory detection in combination with the inventory records to determine a supply priority sequence, and finally optimizing raw material feeding and personnel allocation according to the supply priority and the resource allocation plan, generating personnel scheduling instruction combinations, performing running load checking in combination with the state monitoring information, and outputting a final resource allocation execution scheme to realize efficient scheduling and optimization of production resources.

Description

Technical Field

[0001] This invention relates to the field of production line information control technology, and in particular to a production line information control method and system based on the Internet of Things. Background Technology

[0002] Currently, information-based control of production lines is a crucial link in the intelligent and refined development of the manufacturing industry. Its core objective is to utilize Internet of Things (IoT) technology to achieve real-time collection and dynamic scheduling of multi-source data such as equipment status, raw material inventory, and production progress. In existing technologies, Narrowband Internet of Things (NB-IoT) is commonly used as the basic communication method, connecting sensors and control units distributed along the production line to the network to achieve remote data reporting and centralized management. NB-IoT features low power consumption, wide coverage, and massive connectivity, making it suitable for large-scale equipment data collection in complex workshop environments. However, this existing technology still faces challenges in application, including difficulties in integrating heterogeneous data, insufficient real-time processing capabilities, and poor adaptability of dynamic allocation models.

[0003] In actual production, although NB-IoT can stably transmit data on device operating status, inventory, and production progress, the inconsistent protocol standards among different device manufacturers result in significant differences in data formats. This necessitates multiple rounds of format conversion and verification, increasing processing latency. Furthermore, existing dynamic resource allocation methods primarily rely on fixed rules or simple threshold judgments, failing to fully leverage the real-time data characteristics of NB-IoT. This leads to inflexible responses in multi-variable decision-making scenarios such as raw material allocation and personnel deployment, potentially resulting in resource waste or bottleneck delays.

[0004] In summary, existing technologies suffer from low precision in production resource scheduling. Summary of the Invention

[0005] This invention provides an information control method and system for production lines based on the Internet of Things (IoT) to solve the problems of difficulty in real-time integration of multi-source heterogeneous data and insufficient dynamic resource allocation and optimization, thereby achieving efficient scheduling and flexible response of production resources.

[0006] Firstly, to address the aforementioned technical problems, this invention provides an IoT-based method for information-based control of production lines, comprising: acquiring equipment status monitoring information, raw material inventory records, and production progress tracking logs; performing format unification processing and merging on the status monitoring information, raw material inventory records, and tracking logs respectively to obtain a basic data group; based on the basic data group, performing multi-source data fusion through sharded storage and parallel processing to obtain a multi-source data set; based on the multi-source data set, detecting update frequency and transmission delay and performing priority sorting to obtain a key data subset; and based on the key data subset, combining it with the tracking logs in the basic data group... Extract the resource demand distribution at bottleneck points to obtain resource demand characteristic records; based on these records and the status monitoring information in the basic data set, perform dynamic resource allocation analysis to obtain a resource allocation plan; based on the resource allocation plan, and combined with the raw material inventory records in the basic data set and the status monitoring information, perform inventory checks to obtain a replenishment priority sequence; based on the replenishment priority sequence and the resource allocation plan, optimize raw material delivery and personnel allocation to obtain a personnel scheduling instruction combination; based on the personnel scheduling instruction combination and the status monitoring information in the basic data set, perform operational load verification to obtain a resource allocation execution scheme.

[0007] Preferably, the step of fusing multi-source data through sharded storage and parallel processing based on the basic data group to obtain a multi-source data set includes: performing hash partitioning based on the basic data group to obtain a set of partitioned data fragments; allocating tasks based on the set of data fragments to obtain a computation task queue; performing access frequency analysis and performing hierarchical storage based on the computation task queue to obtain a hierarchical data fragment set for hierarchical storage; and performing field consistency and timestamp verification based on the hierarchical data fragment set, and completing the fusion to obtain an integrated multi-source data set.

[0008] Preferably, the step of detecting update frequency and transmission delay and performing priority sorting based on the multi-source data set to obtain a key data subset includes: recording timestamps and calculating update frequency and transmission delay based on the multi-source data set to obtain a delay detection result; when the update frequency in the delay detection result exceeds a preset frequency threshold or the transmission delay in the delay detection result exceeds a preset delay threshold, performing priority sorting according to field weight and timestamp order to obtain a high-priority data subset; and determining the priority processing node execution order based on the high-priority data subset to obtain the key data subset.

[0009] Preferably, the step of extracting the resource demand distribution of bottleneck points and obtaining resource demand feature records based on the key data subset and the tracking logs in the basic data group includes: parsing the process timestamps and resource usage information based on the key data subset and the tracking logs in the basic data group to obtain a process analysis dataset; segmenting the process analysis dataset by time period or process type to obtain a bottleneck location candidate set; extracting resource usage details corresponding to each bottleneck point and analyzing the distribution pattern based on the bottleneck location candidate set to obtain resource demand distribution features; and extracting key demand indicators and forming structured records based on the resource demand distribution features and priority information in the key data subset to obtain resource demand feature records.

[0010] Preferably, the step of performing dynamic resource allocation analysis based on the resource demand characteristic records and the status monitoring information in the basic data group to obtain a resource allocation plan includes: analyzing the equipment operating status based on the resource demand characteristic records and the status monitoring information in the basic data group to obtain operating status analysis results; performing high-load equipment judgment operations based on the operating status analysis results to obtain a high-load candidate set; analyzing the resource demand distribution in the high-load candidate set and sorting it according to scheduling priority to generate a resource allocation plan.

[0011] Preferably, the step of performing inventory detection based on the resource allocation plan, combined with the raw material inventory records in the basic data group and the status monitoring information, to obtain a replenishment priority sequence includes: extracting inventory quantity and timestamp information based on the resource allocation plan and the raw material inventory records in the basic data group to generate an inventory status dataset; generating a replenishment request signal when the inventory quantity in the inventory status dataset is lower than a preset inventory threshold; and performing priority evaluation based on the replenishment request signal and the status monitoring information to obtain a replenishment priority sequence.

[0012] Preferably, the step of optimizing raw material delivery and personnel allocation based on the supply priority sequence and the resource allocation plan to obtain a personnel scheduling instruction combination includes: adjusting the raw material delivery timing sequence according to the supply priority sequence and the resource allocation plan to obtain an adjusted delivery plan; evaluating personnel allocation based on preset optimization indicators according to the adjusted delivery plan to obtain personnel ratio adjustment requirements; dynamically adjusting the personnel allocation ratio according to the personnel ratio adjustment requirements to form a new allocation plan to obtain a personnel optimization plan; and extracting an appropriate instruction sequence based on the personnel optimization plan and a preset scheduling instruction generation rule, and calibrating the raw material delivery and personnel allocation to obtain a personnel scheduling instruction combination.

[0013] Preferably, the step of verifying the operating load and obtaining a resource allocation execution plan based on the personnel scheduling instruction combination and the status monitoring information in the basic data group includes: calculating the stable value of the equipment operating load and generating a load anomaly alarm signal based on the personnel scheduling instruction combination and the status monitoring information in the basic data group; evaluating the alarm level and predicting the equipment maintenance cycle based on the load anomaly alarm signal to obtain the maintenance cycle time range; and adjusting the personnel allocation ratio and forming an optimization plan based on the maintenance cycle time range and the personnel scheduling instruction combination to obtain the resource allocation execution plan.

[0014] Secondly, this invention provides an IoT-based production line information control system, comprising: a data acquisition module, used to acquire equipment status monitoring information, raw material inventory records, and production progress tracking logs, and to perform format unification processing and merging on the status monitoring information, raw material inventory records, and tracking logs respectively to obtain a basic data group; a multi-source data module, used to perform multi-source data fusion through sharded storage and parallel processing based on the basic data group to obtain a multi-source data set; a key data module, used to detect update frequency and transmission delay and perform priority sorting based on the multi-source data set to obtain a key data subset; and a resource requirement module, used to extract bottleneck points based on the key data subset and the tracking logs in the basic data group. The system analyzes the distribution of resource demand to obtain resource demand characteristic records. A resource allocation module performs dynamic resource allocation analysis based on these characteristics and status monitoring information from the basic data set to obtain a resource allocation plan. A replenishment judgment module performs inventory checks based on the resource allocation plan, raw material inventory records from the basic data set, and the status monitoring information to obtain a replenishment priority sequence. A personnel scheduling module optimizes raw material delivery and personnel allocation based on the replenishment priority sequence and the resource allocation plan to obtain a personnel scheduling instruction combination. An execution plan module performs operational load verification based on the personnel scheduling instruction combination and the status monitoring information from the basic data set to obtain a resource allocation execution plan.

[0015] Thirdly, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the Internet of Things-based production line information control method described in any one of the above.

[0016] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the Internet of Things-based production line information control method described above.

[0017] Compared with the prior art, the present invention has the following beneficial effects: (1) Effectively enhances the real-time fusion capability of multi-source heterogeneous data. This invention introduces a segmented storage and parallel processing strategy into the information control of the production line. By performing hash partitioning, hierarchical storage, and field consistency verification on equipment status, raw material inventory, and production progress logs, it achieves efficient integration of data from different sources and in different formats. Compared with existing technologies that rely on single-node processing, this solution significantly reduces the latency of data format conversion and synchronization, enabling the system to complete status updates within milliseconds, thereby supporting the real-time decision-making needs of subsequent dynamic allocation algorithms; (2) Achieving dynamic optimization of resource allocation and inventory management. This invention accurately identifies bottlenecks in the production process by combining key data subsets with production progress logs and generates resource demand characteristic records. Then, it prioritizes and dynamically allocates high-load equipment based on status monitoring information. At the same time, by comparing the current inventory status with preset thresholds, it triggers replenishment requests and generates a replenishment priority sequence to ensure that raw material delivery and replenishment resources can be preferentially matched with key production processes. This process forms a closed-loop resource allocation and inventory optimization mechanism, effectively avoiding production stagnation and waste caused by raw material shortages or surpluses. (3) Enhancing the stability and flexible response capability of the production line. After optimizing raw material input and personnel allocation, this invention also introduces an operational load verification step to detect abnormalities in equipment load status and predict maintenance cycles, and dynamically adjusts the personnel ratio based on the results. Through this mechanism, the system can identify potential risks in advance and prevent equipment overload failures, ensuring production continuity. At the same time, personnel scheduling and resource allocation are continuously corrected based on real-time data, enabling rapid response to changing production demands and abnormal situations, significantly improving the flexibility and overall operating efficiency of the production line. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the production line information control method based on the Internet of Things provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the Internet of Things-based production line information control system provided in the second embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Reference Figure 1 The first embodiment of the present invention provides an information control method for production lines based on the Internet of Things, comprising the following steps: S11, acquire equipment status monitoring information, raw material inventory records and production progress tracking logs, and perform format unification processing and merging on the status monitoring information, raw material inventory records and tracking logs respectively to obtain basic data groups; S12, Based on the basic data group, multi-source data fusion is completed through fragmented storage and parallel processing to obtain a multi-source data set; S13, Based on the multi-source data set, detect the update frequency and transmission delay and perform priority sorting to obtain a key data subset; S14. Based on the key data subset and the tracking logs in the basic data group, extract the resource demand distribution of the bottleneck point to obtain the resource demand feature record. S15. Based on the resource demand characteristic records and combined with the status monitoring information in the basic data group, perform dynamic resource allocation analysis to obtain a resource allocation plan. S16. Based on the resource allocation plan, combined with the raw material inventory records in the basic data group and the status monitoring information, inventory detection is performed to obtain a replenishment priority sequence. S17, Based on the supply priority sequence and the resource allocation plan, optimize the raw material delivery and personnel allocation to obtain a combination of personnel scheduling instructions; S18. Based on the personnel scheduling instruction combination and the status monitoring information in the basic data group, the operating load is checked to obtain the resource allocation execution plan.

[0021] In step S11, it is necessary to acquire equipment status monitoring information, raw material inventory records, and production progress tracking logs. The status monitoring information, raw material inventory records, and tracking logs are then processed and merged to obtain a basic data set. In one specific embodiment, the purpose of step S11 is to acquire multi-source operating data from the production line and unify its format to form a basic data set for subsequent analysis and processing. Firstly, during the acquisition of equipment status monitoring information, key equipment on the production line, such as CNC machine tools, stamping machines, and welding machines, are equipped with temperature, vibration, current, and power sensors. These sensors transmit real-time data to a centralized server via a narrowband Internet of Things (NB-IoT) communication module. The transmission protocol can be MQTT or CoAP. The uploaded data is typically presented in JSON or CSV format, recording key parameters such as equipment ID, timestamp, temperature, and power.

[0022] Subsequently, during the acquisition of raw material inventory records, the system connects with the enterprise's ERP or warehouse management system to obtain information on raw material receipts, deliveries, and current balances. The data synchronization cycle can be set to once every 5 minutes. Inventory data is typically stored in tabular format, with fields including raw material ID, name, current inventory quantity, minimum safety stock, and update time. For example, the inventory of steel plate raw materials is 1200 kg, and the safety stock threshold is 800 kg.

[0023] During the acquisition of production progress tracking logs, the Manufacturing Execution System (MES) provides real-time progress logs for each process, recording information including batch number, process name, completion time, and related equipment ID. The data is mostly in the form of text logs or structured forms. For example, a batch of products completed the welding process on July 26, 2025 at 08:12, with the associated equipment number M003.

[0024] After acquiring the three types of data mentioned above, the data from different sources needs to be standardized using a format unification module. This module first establishes a unified field mapping standard, uniformly identifying time as `timestamp`, device as `device_id`, and inventory as `stock_level`. Simultaneously, it converts the time format to the UTC standard ISO8601 format to ensure consistency of timestamps across different systems. Regarding units, temperature is standardized to degrees Celsius, inventory to kilograms, and current to amperes to avoid subsequent calculation errors. Furthermore, the character encoding format needs to be standardized to UTF-8 to eliminate compatibility issues during cross-system transmission.

[0025] Finally, using timestamps and device IDs as primary key indexes, equipment status monitoring information, raw material inventory records, and production progress logs are merged to generate a standardized basic data set. This basic data set can be stored as a database table or a standardized JSON object for subsequent sharded storage, multi-source data fusion, and dynamic allocation analysis. For example, the merged record could contain the operating temperature, current power, raw material inventory, and corresponding production progress status of device M001 at the time "2025-07-26T08:15:00Z," providing a unified data input for subsequent bottleneck identification and resource allocation.

[0026] In step S12, based on the basic data group, multi-source data fusion needs to be completed through sharded storage and parallel processing to obtain a multi-source data set. This includes: performing hash partitioning based on the basic data group to obtain a set of partitioned data fragments; allocating tasks based on the set of data fragments to obtain a computation task queue; performing access frequency analysis and performing hierarchical storage based on the computation task queue to obtain a hierarchical data fragment set; and performing field consistency and timestamp verification based on the hierarchical data fragment set, and completing the fusion to obtain the integrated multi-source data set.

[0027] In a specific embodiment, step S12 aims to utilize sharded storage and parallel processing technology to perform multi-source data fusion processing on the basic data group formed in step S11, obtaining a multi-source data set that can be used for subsequent priority detection and bottleneck analysis. Specifically, this step first performs a hash partitioning operation, rationally sharding the large-scale basic data according to two dimensions: device and time. The system selects the device identifier field (device_id) and the timestamp field (timestamp) in the basic data group as a composite hash key, calculates the hash value using a consistent hashing algorithm, and maps it to 0 to... The system uses a hash space and then performs a modulo operation based on the preset number of shards to determine the shard number. For example, if the basic data group contains approximately 1 million records and the system presets 64 shards, each shard will be allocated an average of approximately 15,000 to 20,000 records. In this way, multiple types of data (status information, inventory information, progress logs) from the same device within the same time period will be divided into the same shard, avoiding delays caused by frequent cross-shard merging during subsequent integration.

[0028] During the task allocation phase, the system generates a set of data fragments based on the hash sharding results and schedules parallel computing tasks in conjunction with the node status of the server cluster. The scheduling process monitors the CPU utilization, memory usage, and I / O load of each computing node in real time. When the node load is below a preset threshold (e.g., CPU utilization below 50%, I / O latency below 20ms), new data fragment tasks are prioritized for allocation. The task queue contains fragment number, data volume, processing priority, and required computing resource information. The scheduler dynamically allocates tasks according to priority and node idleness to achieve balanced processing. For example, in a cluster with 10 nodes, if node 1 and node 2 are under low load, more fragments are prioritized for allocation to these two nodes to improve overall throughput. Subsequently, the system performs access frequency analysis and tiered storage processing on the allocated fragments. Access frequency statistics are achieved by recording the number of times a fragment is accessed within a specified time window. Fragments with access counts exceeding a set threshold (e.g., more than 100 accesses per minute) are marked as high-frequency fragments and stored on high-speed media such as NVMe SSDs; fragments below the threshold are stored on high-capacity SATA hard drives or distributed object storage systems. This tiered storage strategy reduces the cost of high-speed media while ensuring that high-frequency data can be accessed quickly within milliseconds. For example, device status monitoring information is typically sampled at a second-level frequency, and its access frequency is much higher than that of historical inventory data. Therefore, it is automatically stored on SSDs to support real-time analysis, while inventory data is stored on low-speed media and only accessed when making replenishment decisions.

[0029] Finally, the system performs field consistency and timestamp verification on the hierarchically stored fragments to complete the final fusion. Field consistency verification is achieved through a predefined mapping table, unifying field names from different systems to standard fields. For example, "temp," "temperature," and "temperature" are all mapped to the unified field "temperature," and the unit of the field is verified (e.g., temperature is unified to degrees Celsius, and inventory quantity is unified to kilograms). Timestamp verification compares the time fields of each data source, setting a time tolerance window of ±100 milliseconds. Records exceeding this window are marked and delayed. The delay mechanism involves temporarily storing abnormal records in a buffer and retrying the comparison within the next three data synchronization cycles. If the timestamp still cannot match with other data fragments during the retry period, it is determined to be expired data and marked for discarding to ensure the timeliness and accuracy of the fused data. The verified fragments are aggregated according to device ID and timestamp to generate the final multi-source data set, formatted uniformly as JSON or a structured database table, serving as direct input for subsequent priority sorting and bottleneck analysis. It should be noted that, in another specific embodiment, a production line generates approximately 1.2 million basic data records during a single shift, including 800,000 equipment status monitoring data records, 200,000 raw material inventory records, and 200,000 production progress tracking logs. The system first performs hash partitioning on these 1.2 million records, using a string concatenated with the device ID and timestamp as the composite key input to the consistent hash function H(x)=(a*x+b)modp, where parameters a=131, b=7, and p=2³². After calculating the hash value, the data is partitioned into shards using a modulo operation of 64 shards. Calculations show that each shard is allocated an average of approximately 18,000 records. The 20,000 data records generated by high-frequency equipment such as M001 within 8 hours are all concentrated in the 5th shard.

[0030] After sharding, the system allocates computation tasks based on node status. Assuming the cluster has 10 nodes, with nodes 1 and 2 having a CPU utilization of 30% and the remaining nodes all above 70%, the scheduler prioritizes assigning tasks from shards 1 to 20 to nodes 1 and 2, marking them as high-priority in the task queue. The remaining 44 shards are then evenly distributed among the other nodes. This strategy ensures balanced task allocation and avoids latency on high-load nodes.

[0031] In access frequency analysis, the system sets a threshold of over 100 accesses per minute as high-frequency segments. Statistical analysis showed that segments 5, 7, and 12 were identified as high-frequency segments and stored on NVMe SSDs, while the remaining segments were stored on SATA hard drives. The system then performs field consistency and timestamp verification on the high-frequency segments, unifying the different source fields "temp," "temperature," and "temperature" to "temperature," and standardizing the time field to UTC format, such as "2025-07-26T08:15:00Z." A 100-millisecond tolerance is set for timestamp comparison; if an inventory record's time is 150 milliseconds later than the status monitoring information, it is marked as delayed data and cached separately for synchronization. Finally, after consistency verification and aggregation, the generated multi-source data set contains 1.2 million records with unified fields, stored as a JSON file, and imported into the subsequent priority detection module, providing complete data input for bottleneck analysis and dynamic resource allocation.

[0032] In step S13, it is necessary to detect the update frequency and transmission delay based on the multi-source data set and perform priority sorting to obtain a key data subset. This includes: recording timestamps and calculating the update frequency and transmission delay based on the multi-source data set to obtain a delay detection result; when the update frequency in the delay detection result exceeds a preset frequency threshold or the transmission delay in the delay detection result exceeds a preset delay threshold, priority sorting is performed according to field weight and timestamp order to obtain a high-priority data subset; based on the high-priority data subset, the execution order of nodes to be processed first is determined to obtain the key data subset.

[0033] In a specific embodiment, the purpose of step S13 is to perform real-time performance testing and priority ranking on the multi-source data set obtained in step S12, so as to quickly identify key data and perform bottleneck analysis. This step includes four stages: timestamp recording and indicator calculation, threshold comparison and anomaly detection, priority ranking, and key data subset generation.

[0034] First, during the timestamp recording and metric calculation phase, the system extracts two types of time information for each data record in the multi-source dataset: one is the timestamp when the data is generated on the device (generation time), and the other is the timestamp when the data enters the fusion database or cache queue (reception time). Both timestamps are uniformly converted to the ISO 8601 standard in UTC format, such as "2025-07-26T08:00:00.123Z", and stored with millisecond precision. The system calculates the transmission latency by calculating the difference between the reception time and the generation time. Simultaneously, it calculates the update frequency based on the number of data arrivals from the same device within a unit time window (e.g., 1 second or 1 minute). The formula for calculating the update frequency is f=N / Δt, where N is the number of records in the window and Δt is the window length. For example, if a device receives 120 records in 1 minute, its update frequency is 2Hz.

[0035] The system then compares the calculated results with preset thresholds. These preset frequency and latency thresholds are determined by production processes and network conditions; for example, the frequency threshold might be set to 1 Hz and the latency threshold to 500 milliseconds. When a device's data update frequency is detected to be below 1 Hz or its transmission latency exceeds 500 milliseconds, the data is deemed to have a real-time anomaly and enters a priority processing flow. Records that simultaneously meet both anomaly criteria are given higher priority, and an anomaly type label is recorded for subsequent analysis.

[0036] During the priority ranking phase, the system scores the records that pass the inspection. The scoring algorithm combines field weights with a time factor. Field weights are pre-set based on production needs; for example, the fault indicator field has a weight of 0.5, the temperature field has a weight of 0.3, and the inventory field has a weight of 0.2. The scoring formula is Score = Σ(weight_i × normalized_value_i). Here, weight_i represents the weight coefficient of the i-th field, determined by the production process's focus on that field, ranging from 0 to 1, and the sum of all field weights is 1. For example, a fault indicator field weight of 0.5 indicates that this field is the most important, accounting for half of the overall score. normalized_value_i represents the normalized value of the i-th field. Normalization transforms the original data value to 0-1 through a linear mapping; the inventory field is mapped to 0-1 within the range of 0 to 1000 kg. Σ represents the weighted summation of all selected fields, and the result is the total score of the data record, reflecting its overall priority, denoted as the priority coefficient. To avoid errors caused by inconsistent units across different fields, the system first normalizes the original values ​​(e.g., to the range of 0 to 1). Furthermore, under the same scoring conditions, data with timestamps closer to the current moment has higher priority, enabling timely responses to critical anomalies.

[0037] Finally, the system selects a high-priority subset of data based on the sorting results and allocates it according to the current load of each node in the cluster. The system will distribute the top 10%-20% of high-priority data to nodes with a load below 50% to ensure that the bottleneck analysis and resource allocation modules prioritize processing critical data. The remaining ordinary data will be processed in a conventional queue for subsequent analysis.

[0038] For example, taking a car manufacturing production line as an example, the multi-source dataset contains 500,000 records, with a sampling frequency of 20 records per second. The system calculates the transmission latency for each record. For instance, if a device's data generation time is "2025-07-26T08:00:00.100Z" and its reception time is "2025-07-26T08:00:00.450Z", the transmission latency is 350 milliseconds. Simultaneously, the system counts 120 records received by this device within one minute, with an update frequency of 2Hz. The system sets a frequency threshold of 1Hz and a latency threshold of 500 milliseconds, determining that the device's data real-time performance is normal. Another device's data frequency is only 0.5Hz, which is considered a low-frequency anomaly and enters the priority sorting process. During sorting, if the device simultaneously exhibits an abnormal temperature field (above 80℃, weight 0.3) and an inventory warning (below the threshold, weight 0.2), the total score is 0.5. After sorting by timestamp, this record is placed at the top of the high-priority queue. Ultimately, the system allocates a subset of high-priority data to three idle computing nodes for processing, enabling critical data to complete bottleneck identification within 200 milliseconds, while ordinary data is processed later.

[0039] It should be noted that the preset frequency threshold is determined by the real-time requirements of the production line. For critical process equipment, such as welding machines and stamping machines, their status information needs to be updated at least once per second to ensure that faults or performance fluctuations can be captured in a timely manner. Therefore, the frequency threshold can be set to 1Hz. For relatively non-critical process equipment, such as finished product packaging machines, the required update frequency of their status data is lower, and the threshold can be set to 0.2Hz (i.e., once every 5 seconds). In practical applications, the system can maintain a frequency threshold table according to equipment type and compare the current update frequency with the corresponding threshold in real time to achieve differentiated monitoring. For example, when the welding machine's update frequency is below 1Hz, high-priority sorting is triggered immediately, while sorting is only triggered when the packaging machine's update frequency is below 0.2Hz, avoiding false alarms from irrelevant equipment.

[0040] It should be noted that the preset latency threshold is set based on network bandwidth and service tolerance. For real-time control processes requiring millisecond-level response, such as temperature over-limit protection or emergency shutdown commands, the latency threshold can be set to 100-200 milliseconds; for inventory and progress data that allow for slight delays, the latency threshold can be appropriately relaxed to 500 milliseconds. In a typical scenario, the average transmission latency of the NB-IoT network deployed in a production workshop is approximately 250 milliseconds. The system sets the threshold for highly sensitive data (such as fault alarms) at 200 milliseconds, and the threshold for low-sensitivity data (such as inventory updates) at 500 milliseconds, thereby achieving classification and judgment. When a temperature sensor data latency exceeding 200 milliseconds is detected, the system immediately triggers high-priority sorting and prioritizes the allocation of node resources, while a 300-millisecond latency in inventory data does not trigger a warning.

[0041] In step S14, based on the key data subset and the tracking logs in the basic data group, it is necessary to extract the resource demand distribution of bottleneck points to obtain resource demand feature records. This includes: parsing the process timestamps and resource usage information based on the key data subset and the tracking logs in the basic data group to obtain a process analysis dataset; segmenting the process analysis dataset by time period or process type to obtain a bottleneck location candidate set; extracting resource usage details corresponding to each bottleneck point and analyzing the distribution pattern based on the bottleneck location candidate set to obtain resource demand distribution features; and extracting key demand indicators and forming structured records based on the resource demand distribution features and priority information in the key data subset to obtain resource demand feature records.

[0042] First, during the timestamp and resource usage information parsing phase, the system retrieves high-priority equipment status information from the key data subset output in step S13 and performs a field-level join with the production progress log in the basic data group. The join conditions include three key indexes: equipment ID, production batch number, and process number, to ensure that the associated data points to the same production stage. During the join process, if an index conflict is found (e.g., multiple overlapping timestamp records for the same process), the system prioritizes retaining the record with the latest timestamp and marks duplicate records as abnormal for subsequent manual review. Regarding timestamp processing, the system uniformly converts the time formats of different data sources to the UTC standard ISO 8601 format and stores them with millisecond-level precision to ensure temporal consistency for subsequent multi-source data comparisons. If a timestamp is missing or incorrectly formatted, the system marks the record as an "abnormal time record" and temporarily stores it in a buffer, allowing for re-matching within the next two data acquisition cycles. If the error persists, the record is discarded. In the standardized processing of resource usage information, the system establishes a unified field-to-unit mapping table, mapping field names from different sources to standard fields. For example, "temp," "temperature," and "temperature" are uniformly mapped to "temperature," and power (kW) and electricity (Wh) are uniformly converted to energy consumption (kWh), man-hours are uniformly converted to minutes, and raw material consumption is uniformly converted to kilograms. For fields with unknown units or unlabeled units, the system marks them as "units to be confirmed" and does not participate in calculations until manual confirmation or automatic rule supplementation.

[0043] Through the above analysis and standardization process, the system finally generates a process analysis dataset. This dataset contains standardized fields such as process ID, start and end time, raw material consumption, energy consumption, labor hours, and equipment status labels. It is also clearly defined for time slicing and bottleneck location identification in subsequent steps, providing basic input for dynamic resource allocation.

[0044] For example, on a welding production line, the original log entry is "2025 / 07 / 26 08:00 Electricity = 1200Wh Labor = 0.5h". After parsing and processing, a standardized record is generated as "2025-07-26T08:00:00.000Z Electricity = 1.2kWh Labor = 30min", and successfully associated with the key data subset according to the equipment ID and process number. If the log is missing a time field, the system marks the record as abnormal and attempts to complete it after two delayed collection cycles. If a match is still not found, the record is discarded.

[0045] Secondly, in the segmentation stage based on time period and process type, the system divides the process analysis dataset into multiple small segments based on production cycle time or process classification, facilitating localized analysis of bottleneck characteristics. Time segments can be divided by fixed duration, such as every 30 minutes or 1 hour; process segments are divided by process type, such as welding, assembly, painting, etc. The system calculates the peak, average, and standard deviation of resource usage for each segment and compares them with the historical normal values ​​for that process. When the peak or average value of a segment exceeds the normal value by 20% to 30%, that segment is marked as a candidate bottleneck location. For example, if the historical average power consumption of the welding process is 1.2 kWh, while the current segment's power consumption reaches 1.8 kWh, exceeding the limit by 50%, it indicates that the energy supply demand for this process far exceeds the historical average level, and this process is then included in the candidate set.

[0046] Subsequently, in the bottleneck location resource usage details extraction stage, the system performs resource segmentation analysis on each link within the candidate set. Specifically, this includes: raw material consumption rate (raw material consumption per unit time), used to determine material supply pressure; production energy consumption (electricity consumption during production periods or links), used to assess energy supply demand; and labor hour distribution (human concentration during peak periods), used to identify imbalances in human resource allocation.

[0047] The system plots these data as time series curves and extracts key features, such as peak time points, peak amplitude, duration, and fluctuation range, to further determine the nature of the bottleneck (such as periodic peaks or sudden anomalies).

[0048] Finally, in the structured record generation stage, the system combines the above analysis results with the priority information provided in step S13 to prioritize the importance of bottleneck links. The scoring formula is: Priority Score = α × Resource Demand Intensity + β × Duration Weight + γ × Priority Coefficient.

[0049] α, β, and γ are set according to production needs, for example, 0.5, 0.3, and 0.2 respectively. Resource demand intensity is determined by the ratio of current peak consumption to historical average consumption. Duration weight is calculated from the overload duration, and priority coefficient is obtained from priority information. Bottlenecks with a score higher than 0.8 are marked as Level 1 bottlenecks and prioritized for subsequent allocation steps. The final generated structured record contains fields such as: bottleneck ID, time period, resource type (electricity / raw materials / labor), peak demand value, demand change trend, and priority score, and is stored in JSON or database table format.

[0050] For example, on a car body welding production line, after parsing key data subsets and production logs, the system found that welding stage M003 exhibited resource anomalies between 08:00 and 08:30 on July 26, 2025: energy consumption was 1.8 kWh, while the historical average was 1.2 kWh, exceeding the limit by 50%; simultaneously, the manual labor time for this stage was 8 minutes per car body, 6 minutes higher than the normal value, an increase of 33%. After the system performed statistical analysis in 30-minute segments, it included the welding stage during this period in the bottleneck candidate set. Further analysis revealed that this stage experienced periodic power fluctuations throughout the shift (08:00-12:00), with peaks concentrated in the 08:00-08:30 and 10:30-11:00 time periods. Combining the priority coefficient (0.9) from step S13, the system calculated a priority score of 0.88, marking it as a level-one bottleneck, and generating the following structured record example: { "bottleneck_step": "welding", "time_window": "2025-07-26T08:00:00Z-2025-07-26T08:30:00Z", "energy_peak_kWh": 1.8, "labor_peak_min": 8, "priority_score": 0.88, "trend": "periodic peak" } This record is written to the database and used in subsequent dynamic resource allocation steps to prioritize the allocation of electrical and human resources.

[0051] In step S15, dynamic resource allocation analysis is performed based on the resource demand characteristic records and the status monitoring information in the basic data group to obtain a resource allocation plan. This includes: analyzing the equipment operating status based on the resource demand characteristic records and the status monitoring information in the basic data group to obtain operating status analysis results; performing high-load equipment judgment operations based on the operating status analysis results to obtain a high-load candidate set; analyzing the resource demand distribution in the high-load candidate set and sorting it according to scheduling priority to generate a resource allocation plan.

[0052] In a specific embodiment, the objective of step S15 is to comprehensively analyze the operating status of each piece of equipment on the production line based on the resource demand characteristic record generated in step S14 and combined with the status monitoring information in the basic data group, identify high-load equipment and determine the priority allocation order, thereby forming a dynamic resource allocation plan.

[0053] First, in the operational status analysis phase, the system matches resource demand characteristic records with status monitoring information, using device ID and timestamp as association conditions. The system extracts key operating parameters of the equipment, such as real-time power, temperature, vibration amplitude, and runtime, and compares them with peak time periods in the resource demand characteristics to determine whether the equipment is continuously operating during peak periods or shows signs of overload. To improve analysis accuracy, the system uses a sliding window algorithm to smooth the parameter sequence. Specifically, it uses a 5-minute analysis window to calculate the average value and fluctuation range of the parameters within that window. The fluctuation range is calculated using the standard deviation method, which involves taking the square root of the average of the squared deviations of each data point within the window from the average value to obtain the stability index of the parameter within that time period. Significant fluctuations are considered when the standard deviation exceeds a preset threshold (e.g., power standard deviation exceeding 0.2kW or temperature standard deviation exceeding 3℃). Furthermore, in some scenarios, the system can use the range method (the difference between the maximum and minimum values) as an auxiliary means to quickly assess short-term drastic fluctuations, but the standard deviation calculation result is used as the final judgment basis. This method can effectively filter out misjudgments caused by short-term spike signals, ensuring the accuracy of high-load determination.

[0054] It should be noted that, in one specific embodiment, the system pre-sets a judgment rule based on standard deviation for the fluctuation range used in the operational status analysis. Specifically, after processing with the sliding window algorithm, the average value of each data point within the window is calculated, and the sum of the squared deviations of all data points relative to the average value is calculated. The square root of this sum is then taken to obtain the standard deviation value. If this standard deviation exceeds a preset threshold, the operational parameter is determined to have significant fluctuations. For example, for power parameters, a standard deviation exceeding 0.2kW is considered a high-fluctuation state; for temperature parameters, a standard deviation exceeding 3℃ is considered a high-fluctuation state. This threshold can be adjusted in the configuration file according to equipment characteristics and production safety standards. To cope with sudden anomalies, the system also provides an auxiliary mechanism for range judgment, which calculates the difference between the maximum and minimum values ​​within the window. When the range exceeds a preset range threshold (e.g., a power range of 0.5kW), an auxiliary fluctuation flag is triggered. Ultimately, the standard deviation judgment result is the primary factor, while the range result serves as a basis for verification or early warning, thereby improving the robustness and applicability of fluctuation identification.

[0055] Subsequently, during the high-load equipment judgment phase, the system filters the operating status analysis results based on preset thresholds. These preset thresholds include a power utilization rate threshold (e.g., 80% of rated power), a temperature threshold (e.g., 75℃), and a comprehensive load index. The comprehensive load index is calculated by weighting three parameters: power, temperature, and vibration. The formula is Load_Index=α×P_norm+β×T_norm+γ×V_norm, where P_norm is the normalized power utilization rate, T_norm is the normalized temperature, and V_norm is the normalized vibration amplitude. The weights α, β, and γ are set to 0.5, 0.3, and 0.2, respectively, and their sum is 1. When the comprehensive load index exceeds the preset threshold (e.g., 0.75), or when the equipment's power exceeds 80% and its temperature exceeds the threshold for three consecutive window periods, the equipment is included in the high-load candidate set, and its overload duration and peak parameters are recorded. For example, welding machine M003 had an average power of 4.8kW (rated 6kW) between 08:00 and 08:30, a utilization rate of 80%, and a temperature of 85℃. Its comprehensive load index was calculated to be 0.81, exceeding the 0.75 threshold, and it exceeded the standard for three consecutive windows, thus being marked as a high-load device. Next, in the resource demand distribution and scheduling priority ranking stage, the system performs a multi-dimensional analysis of the resource demands of devices within the high-load candidate set, including raw material demand rate, power consumption rate, and labor hours input, and combines this with the priority score generated in step S14 for comprehensive ranking. The ranking algorithm can use a weighted scoring model, for example: Comprehensive score = 0.5 × Resource demand intensity + 0.3 × Duration weight + 0.2 × Priority score.

[0056] Resource demand intensity is determined by the ratio of current peak consumption to historical average consumption, duration weight is calculated based on overload duration, and priority scoring directly references bottleneck characteristic records. Equipment with a comprehensive score higher than 0.8 is designated as a Level 1 allocation target. Based on the range of comprehensive scores, equipment is divided into three levels of allocation targets: Level 1 allocation target: Comprehensive score higher than 0.8 indicates that the equipment is under extremely high resource pressure and requires priority access to raw materials and personnel, with scheduling to be executed immediately in the next production cycle.

[0057] Secondary allocation target: The comprehensive score is between 0.6 and 0.8, indicating that the equipment is under medium to high resource pressure and needs to complete the scheduling optimization within two production cycles. It can be coordinated in parallel with the primary target.

[0058] Level 3 allocation target: If the comprehensive score is below 0.6 but above 0.4, it is listed as a low-priority monitoring target and will only be scheduled after resource surplus or the load of other equipment decreases. If it is below 0.4, it is considered a normal state equipment and will not enter the allocation sequence in this round.

[0059] Finally, during the resource allocation plan generation phase, the system generates an allocation scheme based on the sorting results, including the priority and proportion of raw materials, electricity, and labor. The scheme is stored in a JSON structure, with fields including equipment ID, allocation resource type, allocation proportion, and allocation time period, and is simultaneously sent to the resource scheduling module for execution. For example, welding machine M003 receives a priority allocation of 30% of the electricity reserve and 2 additional operators, while assembly machine M005 receives 20% of the raw material supply.

[0060] For example, taking a car body welding production line as an example, the resource demand characteristic record shows that welding machine M003 has a peak power consumption of 1.8kWh and 8 minutes of manual labor between 8:00 and 8:30, with a priority score of 0.88. Status monitoring information shows that its power utilization rate is 85%, temperature is 82℃, and vibration deviation is 0.05g during this period, all exceeding the threshold. The system determines it as a high-load device and calculates a comprehensive score of 0.9 based on the resource demand distribution, listing it as a first-level allocation target; simultaneously, assembly machine M005 has a power utilization rate of 75% and a 20% increase in manual labor demand, with a score of 0.75, listing it as a second-level allocation target. Based on the ranking results, the system generates an allocation plan: prioritizing the allocation of an additional 30% of power and 2 operators to M003, while allocating 20% ​​of raw material supply to M005. The plan is output in JSON format. { "priority_order": ["M003", "M005"], "allocation": [ {"device_id": "M003", "resource": "power", "allocation": "30%"}, {"device_id": "M003", "resource": "labor", "allocation": "2 people"}, {"device_id": "M005", "resource": "material", "allocation": "20%"} ] } In step S16, inventory detection is performed based on the resource allocation plan, combined with the raw material inventory records in the basic data group and the status monitoring information, to obtain a replenishment priority sequence. This includes: extracting inventory quantity and timestamp information based on the resource allocation plan and the raw material inventory records in the basic data group to generate an inventory status dataset; generating a replenishment request signal when the inventory quantity in the inventory status dataset is lower than a preset inventory threshold; and performing priority evaluation based on the replenishment request signal and the status monitoring information to obtain a replenishment priority sequence.

[0061] First, during the inventory status dataset generation phase, the system matches the target equipment and required raw material types involved in the resource allocation plan with the inventory records in the basic data group. To ensure matching accuracy, the system pre-establishes a raw material ID mapping table, which contains standardized raw material names and their synonym lists. For example, "steel_plate," "steel plate," and "SP-01" are uniformly mapped to the standard field "steel_plate." During matching, the system first performs field standardization, converting the raw material descriptions in the allocation plan and inventory records into a standardized field format (e.g., all lowercase English letters, removing special symbols), and then performs a one-to-one comparison according to the standard fields. If a synonym match is found (e.g., "steel plate" corresponds to "steel_plate"), the system determines it as a valid match according to the mapping table rules; if an unrecognizable field is encountered (e.g., no corresponding synonym), it is marked as an abnormal field, and the record is awaiting manual confirmation or automatic updating of the mapping table before retrying. After a successful match, the system extracts the current inventory quantity, unit, entry time, and last update timestamp of the corresponding raw material, and uniformly converts it to ISO 8601 UTC format for storage. For example, when the resource allocation plan requires a replenishment of 50kg of steel plate for welding machine M003, the system will retrieve the current inventory quantity of the steel plate (e.g., 120kg) and the last update time (e.g., 2025-07-26T08:00:00Z) from the inventory records to form a preliminary inventory status dataset. Secondly, in the inventory threshold comparison and replenishment request generation stage, the system compares the current inventory quantity with a preset inventory threshold. The preset inventory threshold is determined based on the production safety stock strategy; for example, the safety stock of key raw material steel plate can be set to 100kg, and welding wire to 30kg. When the inventory of a certain raw material falls below the corresponding threshold, the system generates a replenishment request signal, along with the raw material type, the quantity in shortfall, and the identification information of the corresponding equipment. If the inventory is normal, it is marked as not requiring replenishment.

[0062] Next, in the priority assessment phase, the system combines supply request signals with status monitoring information for analysis, assessing the urgency of supply needs for different equipment. Assessment indicators include: current equipment load rate (e.g., power utilization ≥ 80%), inventory gap ratio (the ratio of actual gap to safety stock threshold), and the comprehensive score generated in step S15, the dynamic resource allocation analysis phase. The system uses a weighted calculation formula to score the three indicators, for example: total score = 0.4 × load rate + 0.3 × inventory gap ratio + 0.3 × comprehensive score, and ranks the equipment based on the total score. The ranking results are used to generate a supply allocation plan, which clarifies the allocation order and proportion of raw materials, electricity, and labor, and reserves more resource quotas for high-scoring equipment based on priority scores, ensuring that critical equipment is prioritized and thus optimizing the overall production rhythm. The supply priority value for each piece of equipment is calculated and arranged in descending order to form a supply priority sequence.

[0063] It should be noted that in a certain production workshop, the resource allocation plan requires the supply of 50kg of steel plate to welding machine M003 and 30kg of aluminum plate to assembly machine M005. Inventory records show that the current steel plate inventory is 120kg (safety stock threshold 100kg), and the aluminum plate inventory is 40kg (safety stock threshold 50kg). The system calculation shows that the steel plate inventory is within the safe range, so no replenishment request is triggered; the aluminum plate inventory is below the threshold, with a shortage of 10kg, therefore a replenishment request signal is generated for the aluminum plate. Combining the status monitoring information, the current power utilization rate of assembly machine M005 is 78%, and its priority score in the allocation plan is 0.7. The system calculates the overall priority as: 0.5×0.78+0.3×0.2+0.2×0.7=0.59.

[0064] After sorting, this request is classified as a medium-priority supply task. The system ultimately generates a supply priority sequence, output as a JSON structure example: { "priority_order": ["M005"], "requests": [ {"device_id": "M005", "material": "aluminum_plate", "gap": "10 kg", "priority_score": 0.556} ] } This priority sequence serves as the input for the next step, S17, optimizing raw material input and personnel allocation.

[0065] It should be noted that the preset inventory threshold is determined based on the production line's safety stock strategy and historical consumption data. The system first calculates the average daily consumption of each raw material over the past three months and sets a safety factor, typically 1.2 to 1.5 times, based on the production cycle fluctuation coefficient. For example, for steel plate raw materials, the historical average daily consumption is 100 kg with a fluctuation coefficient of 1.2, so the safety stock threshold is set at 120 kg; for welding wire raw materials, the average daily consumption is 20 kg with a fluctuation coefficient of 1.5, so the threshold is set at 30 kg. This threshold is stored in the configuration database and maintained as an independent record by raw material ID, allowing for differentiated configuration for different equipment needs. When the actual inventory level falls below the corresponding threshold, the system automatically triggers a replenishment request and records the gap value and corresponding equipment identifier for priority evaluation.

[0066] In step S17, it is necessary to optimize the raw material delivery and personnel allocation based on the supply priority sequence and the resource allocation plan to obtain a personnel scheduling instruction combination. This includes: adjusting the raw material delivery timing sequence according to the supply priority sequence and the resource allocation plan to obtain an adjusted delivery plan; evaluating personnel allocation based on preset optimization indicators according to the adjusted delivery plan to obtain personnel ratio adjustment requirements; dynamically adjusting the personnel allocation ratio according to the personnel ratio adjustment requirements to form a new allocation plan to obtain a personnel optimization plan; and extracting an adaptive instruction sequence based on the personnel optimization plan and a preset scheduling instruction generation rule, and calibrating the raw material delivery and personnel allocation to obtain a personnel scheduling instruction combination.

[0067] First, during the raw material delivery timing adjustment phase, the system matches the priority scores in the replenishment priority sequence with the raw material demand in the resource allocation plan to generate a list of priority raw materials. The delivery timing is then rearranged using a time-series algorithm. For example, raw materials with a priority score higher than 0.8 are scheduled for delivery in the previous cycle of the current production cycle, while raw materials with a priority score between 0.5 and 0.8 are scheduled for delivery in the next cycle to avoid bottlenecks caused by delivery delays. For instance, when the steel plate replenishment task score for welding machine M003 is 0.9, its delivery timing is advanced to the next 10-minute cycle of the current shift.

[0068] Subsequently, during the personnel ratio adjustment needs assessment phase, the system evaluates the rationality of personnel allocation based on the adjusted delivery plan. Assessment indicators include: raw material delivery volume per unit time, manual operation capacity (delivery volume per person per minute), and the parallel requirements of different processes. For example, if the raw material delivery task increases by 30%, and the existing personnel load has reached 85%, the system determines that the personnel ratio needs to be increased, such as expanding the current 5-person delivery team to 7 people to maintain the delivery rhythm.

[0069] Next, in the dynamic adjustment phase of personnel allocation, the system generates new personnel optimization plans based on the evaluation results. Adjustment methods include cross-process reassignment (temporarily transferring personnel from low-load processes to high-load processes) and dynamic shift work (increasing staff shifts during peak hours). The system sets maximum and minimum manpower boundaries for each process to avoid excessive concentration or dispersion. For example, if the welding process allows a maximum of 8 operators, and the optimization plan requires 9, the system automatically prompts for additional temporary personnel or adjustments to the deployment rhythm.

[0070] Finally, in the dispatch instruction generation and calibration phase, the system integrates the personnel optimization plan with the raw material delivery plan according to preset instruction generation rules to generate an executable instruction sequence. The instruction content includes: delivery start and end times, corresponding equipment IDs, type and quantity of raw materials to be delivered, personnel assignment ratios, and task switching times. The calibration phase verifies the feasibility of the instructions by comparing them with real-time status monitoring information (such as delivery completion confirmation signals and personnel attendance status), and corrects the instructions in real time based on anomalies (such as personnel absence or raw material shortages). Finally, the system outputs a combination of personnel dispatch instructions and sends it to the execution terminal.

[0071] For example, on a car body welding production line, the supply priority sequence shows that the welding machine's M003 steel plate delivery task score is 0.9, and the assembly machine's M005 aluminum plate delivery task score is 0.6. The system prioritizes adjusting the delivery timing of the M003 steel plate to 10 minutes earlier than the current cycle, and postpones the delivery of the M005 aluminum plate to the next cycle. After the adjustment, the total delivery task increases by 20% compared to the original plan, and the average load of the existing 5-person delivery team increases to 95%, exceeding the preset 80% threshold. The system determines that 2 more people are needed to assist the welding process, and 1 person is transferred from the low-load packaging process to fill the gap. The optimized personnel configuration is 7 people in the welding team and 4 people in the assembly team.

[0072] It should be noted that, in one specific embodiment, a preset optimization index is used to evaluate the degree to which the personnel allocation plan matches the raw material delivery efficiency and production cycle stability. This index consists of three types of parameters: task matching rate, personnel load balance, and cycle time achievement rate. The task matching rate measures the degree of matching between the number of personnel and the amount of tasks to be delivered, and its calculation formula is the ratio of the actual available delivery quantity to the required delivery quantity, with a target set at ≥95%. The personnel load balance measures the uniformity of personnel distribution among different processes, calculated through the standard deviation of the load rate of each process, with a target set at ≤10%. The cycle time achievement rate measures whether the delivery of key links can be completed within the specified time after adjustment, with a target set at ≥90%. When the comprehensive optimization index is lower than the preset value (e.g., 80 points), the system automatically triggers cross-process scheduling or shift optimization to ensure that key tasks are prioritized.

[0073] It should be noted that the preset scheduling instruction generation rules are used to transform the optimized raw material delivery and personnel allocation results into standardized executable instructions. These rules define the instruction structure, field priorities, and exception handling mechanisms. The instruction structure includes core fields such as equipment number, raw material type and quantity, delivery start and end times, number of participating personnel, and role assignments. Field priorities follow the sorting logic of "equipment → raw material → time → personnel" to ensure that the executor prioritizes parsing key parameters. The exception handling mechanism requires the automatic generation of emergency instructions when the delivery time is approaching and personnel are not yet in place or raw materials have not arrived, postponing the task to the next cycle or triggering the call of backup personnel. Instructions are uniformly output in JSON format and synchronized to the MES system and on-site dashboards via API interfaces to support real-time monitoring and manual intervention.

[0074] In step S18, based on the personnel scheduling instruction combination and the status monitoring information in the basic data group, it is necessary to perform an operational load verification to obtain a resource allocation execution plan. This includes: calculating the stable value of the equipment operating load and generating a load anomaly alarm signal based on the personnel scheduling instruction combination and the status monitoring information in the basic data group; evaluating the alarm level and predicting the equipment maintenance cycle based on the load anomaly alarm signal to obtain the maintenance cycle time range; and adjusting the personnel allocation ratio and forming an optimization plan based on the maintenance cycle time range and the personnel scheduling instruction combination to obtain the resource allocation execution plan.

[0075] First, the system combines the personnel dispatch instructions generated in step S17 with the status monitoring information in the basic data group. Data alignment is performed using the equipment number and timestamp field to extract real-time operating parameters such as power, speed, temperature, and vibration amplitude for each device. A sliding window method (e.g., a 5-minute window) is used to smooth the collected data, and a stable operating load value is calculated within the window. The stable value is mathematically defined as the arithmetic mean of the collected points within the time window, combined with the standard deviation to assess the degree of fluctuation: Stable value = Σ(parameter value) / N, where N is the number of sampling points within the window; simultaneously, the standard deviation = sqrt(Σ(parameter value − stable value)² / N) is calculated to aid in determining stability. When the stable value exceeds a preset load threshold (e.g., 80% of rated power or a temperature threshold of 75℃), it is determined to be a high-load state; if the stable value does not exceed the threshold but the standard deviation exceeds a preset fluctuation threshold (e.g., power fluctuation > 0.2kW), it is marked as an abnormal fluctuation state. When generating an abnormal load alarm signal, the system will also include the device ID, the type of abnormality (power, temperature, or a combination of both), and the extent of the exceedance, in order to predict subsequent maintenance cycles and optimize personnel scheduling.

[0076] Subsequently, the system assesses the alarm level of the load anomaly alarm signals, classifying them into three levels based on the degree of exceedance and duration: Level 1 alarms are severe overloads (exceeding the threshold by more than 30% and lasting for more than 10 minutes), Level 2 alarms are moderate overloads (exceeding the threshold by 10%-30% and lasting for more than 5 minutes), and Level 3 alarms are minor overloads (just exceeding the threshold and lasting for less than 5 minutes). Based on the alarm level, the system retrieves historical maintenance records and operating condition data, and uses a decision tree method to predict the time range of the next maintenance cycle for the equipment, for example, predicting that preventive maintenance is required within the next 15-20 days.

[0077] Finally, the system dynamically adjusts the personnel allocation ratio and forms an optimized plan by combining the predicted maintenance cycle time range with personnel scheduling instructions. For equipment with level one or two alarms, the system prioritizes increasing the proportion of inspection and maintenance personnel and reduces the amount of raw materials supplied during high-load periods to reduce equipment pressure. For equipment with level three alarms, the system only increases the frequency of short-term inspections within the planned maintenance cycle without changing the supply plan. The optimized resource allocation execution plan includes fields such as equipment priority, resource supply ratio, personnel allocation plan, and maintenance time window, and is output to the execution terminal in a standardized format to achieve coordinated control of raw material supply, personnel allocation, and maintenance plan.

[0078] For example, taking a car body welding production line, a personnel scheduling instruction requires 50kg of steel plate to be fed into welding machine M003 at 8:10 AM and 7 people to operate it. Status monitoring data shows that from 8:00 to 8:10 AM, M003's power stability was 5.2kW (rated 6kW), with a utilization rate of 87%; the temperature was 82℃, exceeding the 75℃ threshold. The system determined this to be a level two alarm and predicted that the equipment would require maintenance within the next 18 days. Based on this, the system adjusted the personnel allocation for the subsequent three shifts of M003 from 7 people to 5 operators + 2 inspectors, while reducing the raw material feed rate by 10% per shift to ensure reduced equipment pressure before maintenance.

[0079] In summary, this invention provides an information control method and system for production lines based on the Internet of Things (IoT) to solve the problems of difficulty in real-time integration of multi-source heterogeneous data and insufficient dynamic resource allocation and optimization, thereby achieving efficient scheduling and flexible response of production resources.

[0080] Reference Figure 2The second embodiment of the present invention provides an IoT-based production line information control system, comprising: a data acquisition module, used to acquire equipment status monitoring information, raw material inventory records, and production progress tracking logs, and to perform format unification processing and merging on the status monitoring information, raw material inventory records, and tracking logs respectively to obtain a basic data group; a multi-source data module, used to perform multi-source data fusion through sharded storage and parallel processing based on the basic data group to obtain a multi-source data set; a key data module, used to detect update frequency and transmission delay and perform priority sorting based on the multi-source data set to obtain a key data subset; and a resource requirement module, used to extract bottleneck points based on the key data subset and the tracking logs in the basic data group. The system analyzes the distribution of resource demand to obtain resource demand characteristic records. A resource allocation module performs dynamic resource allocation analysis based on these characteristics and status monitoring information from the basic data set to obtain a resource allocation plan. A replenishment judgment module performs inventory checks based on the resource allocation plan, raw material inventory records from the basic data set, and the status monitoring information to obtain a replenishment priority sequence. A personnel scheduling module optimizes raw material delivery and personnel allocation based on the replenishment priority sequence and the resource allocation plan to obtain a personnel scheduling instruction combination. An execution plan module performs operational load verification based on the personnel scheduling instruction combination and the status monitoring information from the basic data set to obtain a resource allocation execution plan.

[0081] It should be noted that the IoT-based production line information control system provided in this embodiment of the invention is used to execute all process steps of the IoT-based production line information control method in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0082] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a resource allocation program. When the processor executes the computer program, it implements the steps described in the various embodiments of the IoT-based production line information control method, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above-described device embodiments, such as the resource allocation module.

[0083] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0084] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0085] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0086] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0087] Wherein, if the modules / units integrated in the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0088] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0089] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A production line information control method based on the Internet of Things, characterized in that... include: Acquire equipment status monitoring information, raw material inventory records, and production progress tracking logs, and perform format unification processing and merging on the status monitoring information, raw material inventory records, and tracking logs respectively to obtain basic data groups; Based on the aforementioned basic data set, multi-source data fusion is achieved through sharded storage and parallel processing to obtain a multi-source data set. Based on the multi-source data set, the update frequency and transmission delay are detected and priority sorting is performed to obtain a key data subset. This includes: recording timestamps and calculating update frequency and transmission delay based on the multi-source data set, including: extracting two types of time information for each data record in the multi-source data set: one is the timestamp when the data is generated on the device, i.e., the generation time; the other is the timestamp when the data enters the fusion database or cache queue, i.e., the reception time. The transmission delay is obtained by calculating the difference between the reception time and the generation time. At the same time, the update frequency is calculated based on the number of times data arrives from the same device within a unit time window to obtain the delay detection result. When the update frequency of the data in the delay detection result is lower than a preset frequency threshold... If the transmission delay of the data in the delay detection result exceeds a preset delay threshold, it is determined that the data has a real-time anomaly and enters the priority processing flow. The data with real-time anomalies is prioritized according to field weight and timestamp order, including: calculating a score for records that pass the detection, where the scoring algorithm combines field weight and time factors. The field weight is preset according to production needs, and the result is the total score of the data record, denoted as the priority coefficient. Under the same score, data with timestamps closer to the current time has higher priority. A high-priority data subset is selected based on the sorting result. The execution order of priority processing nodes is determined based on the high-priority data subset. Obtain a key data subset; based on the key data subset and the tracking logs in the basic data set, extract the resource demand distribution of bottleneck points to obtain resource demand feature records, including: based on the key data subset and the tracking logs in the basic data set, parse the process timestamps and resource usage information to obtain a process analysis dataset; based on the process analysis dataset, segment the data by time period or process type, calculate the peak, average, and standard deviation of resource usage for each segment, and compare them with the historical normal values ​​for that segment; when the peak or average value of a segment exceeds the normal value by 20% to 30%, mark the segment as a bottleneck location candidate to obtain a bottleneck location candidate set; based on the bottleneck location candidate set, propose... Resource usage details corresponding to each bottleneck point are obtained and their distribution patterns are analyzed to obtain resource demand distribution characteristics. Based on the resource demand distribution characteristics and the priority information in the key data subset, key demand indicators are extracted and structured records are formed to obtain resource demand characteristic records. Based on the resource demand characteristic records and the status monitoring information in the basic data set, dynamic resource allocation analysis is performed to obtain a resource allocation plan. Based on the resource allocation plan and the raw material inventory records in the basic data set and the status monitoring information, inventory checks are performed to obtain a replenishment priority sequence. Based on the replenishment priority sequence and the resource allocation plan, raw material delivery and personnel allocation are optimized to obtain a personnel scheduling instruction combination.Based on the personnel dispatch instruction combination and the status monitoring information in the basic data group, the operational load is checked to obtain a resource allocation execution plan.

2. The IoT-based production line information control method according to claim 1, characterized in that, The step of fusing multi-source data based on the basic data set through sharded storage and parallel processing to obtain a multi-source data set includes: performing hash partitioning based on the basic data set to obtain a set of partitioned data fragments; allocating tasks based on the set of data fragments to obtain a computation task queue; performing access frequency analysis and performing hierarchical storage based on the computation task queue to obtain a hierarchical data fragment set for hierarchical storage; and performing field consistency and timestamp verification based on the hierarchical data fragment set, and completing the fusion to obtain an integrated multi-source data set.

3. The IoT-based production line information control method according to claim 1, characterized in that, The step of performing dynamic resource allocation analysis based on the resource demand characteristic records and the status monitoring information in the basic data group to obtain a resource allocation plan includes: analyzing the equipment operating status based on the resource demand characteristic records and the status monitoring information in the basic data group to obtain operating status analysis results; performing high-load equipment judgment operations based on the operating status analysis results to obtain a high-load candidate set; analyzing the resource demand distribution in the high-load candidate set and sorting it according to scheduling priority to generate a resource allocation plan.

4. The IoT-based production line information control method according to claim 1, characterized in that, The step of performing inventory detection and obtaining a replenishment priority sequence based on the resource allocation plan, combined with the raw material inventory records in the basic data group and the status monitoring information, includes: extracting inventory quantity and timestamp information based on the resource allocation plan and the raw material inventory records in the basic data group to generate an inventory status dataset; generating a replenishment request signal when the inventory quantity in the inventory status dataset is lower than a preset inventory threshold; and performing priority evaluation based on the replenishment request signal and the status monitoring information to obtain a replenishment priority sequence.

5. The IoT-based production line information control method according to claim 1, characterized in that, The step of optimizing raw material delivery and personnel allocation based on the supply priority sequence and the resource allocation plan to obtain a personnel scheduling instruction combination includes: adjusting the timing sequence of raw material delivery according to the supply priority sequence and the resource allocation plan to obtain an adjusted delivery plan; evaluating personnel allocation based on preset optimization indicators according to the adjusted delivery plan to obtain personnel ratio adjustment requirements; dynamically adjusting the personnel allocation ratio according to the personnel ratio adjustment requirements to form a new allocation plan to obtain a personnel optimization plan; and extracting an appropriate instruction sequence based on the personnel optimization plan and preset scheduling instruction generation rules, and calibrating the raw material delivery and personnel allocation to obtain a personnel scheduling instruction combination.

6. The IoT-based production line information control method according to claim 1, characterized in that, The step of verifying the operating load based on the personnel dispatch instruction combination and the status monitoring information in the basic data group to obtain a resource allocation execution plan includes: calculating the stable value of the equipment operating load and generating a load anomaly alarm signal based on the personnel dispatch instruction combination and the status monitoring information in the basic data group; evaluating the alarm level and predicting the equipment maintenance cycle based on the load anomaly alarm signal to obtain the maintenance cycle time range; and adjusting the personnel allocation ratio and forming an optimization plan based on the maintenance cycle time range and the personnel dispatch instruction combination to obtain the resource allocation execution plan.

7. An information control system for production lines based on the Internet of Things, characterized in that, The method for implementing the IoT-based production line information control method according to any one of claims 1-6 comprises: a data acquisition module, used to acquire equipment status monitoring information, raw material inventory records, and production progress tracking logs, and to perform format unification processing and merging on the status monitoring information, the raw material inventory records, and the tracking logs respectively to obtain a basic data group; a multi-source data module, used to perform multi-source data fusion through sharded storage and parallel processing based on the basic data group to obtain a multi-source data set; a key data module, used to detect update frequency and transmission delay and perform priority sorting based on the multi-source data set to obtain a key data subset; and a resource requirement module, used to extract bottle information based on the key data subset and the tracking logs in the basic data group. The system analyzes the distribution of resource demand at key points to obtain resource demand characteristic records. A resource allocation module performs dynamic resource allocation analysis based on these characteristics, combined with status monitoring information from the basic data set, to obtain a resource allocation plan. A replenishment judgment module performs inventory checks based on the resource allocation plan, combined with raw material inventory records from the basic data set and the status monitoring information, to obtain a replenishment priority sequence. A personnel scheduling module optimizes raw material delivery and personnel allocation based on the replenishment priority sequence and the resource allocation plan, to obtain a personnel scheduling instruction combination. An execution plan module performs operational load verification based on the personnel scheduling instruction combination and the status monitoring information from the basic data set, to obtain a resource allocation execution plan.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the Internet of Things-based production line information control method as described in any one of claims 1 to 6.

Images