A collaborative processing method and system of industrial equipment data acquisition and intelligent monitoring board

Abstract

The application provides a kind of industrial equipment data acquisition and intelligent monitoring board cooperative processing method and system, this method includes: the operation data of industrial equipment is executed edge layer coding processing, generates the sampling coding stream containing multiple levels of sampling granularity, and the rendering computing power pool fragmentation quantization processing of each view partition of intelligent monitoring board is executed, generates partition computing power quota image;Based on the partition computing power quota image, the layered coding contract of the acquisition end is generated reversely, the abnormal pre-detection signal uploaded by the acquisition end is received synchronously and the computing power quota of corresponding board partition is reserved for it, to establish the cooperative contract link of sampling coding and board computing power two-way constraint;Generate the effective alarm set consistent with check, and according to the effective alarm set, the computing power fragmentation boundary of each view partition of board is dynamically redrawn;Trace encrypted data stream is mapped to the trace exclusive view partition of board, to form the trace link of sampling granularity and view rendering linkage. The application can greatly improve the cooperative processing efficiency.

Description

Technical Field

[0001] This invention relates to the field of industrial Internet of Things (IoT) technology, and in particular to a collaborative processing method and system for industrial equipment data acquisition and intelligent monitoring dashboards. Background Technology

[0002] With the deep integration of Industrial Internet of Things (IIoT) technology into production scenarios, industrial equipment data acquisition and intelligent monitoring dashboards have become core supports for equipment operation and maintenance and production management. Existing technologies generally adopt a layered, decoupled deployment architecture consisting of a data acquisition layer, a platform layer, and a dashboard layer: the data acquisition layer connects to heterogeneous devices through an industrial gateway to complete raw data acquisition and local threshold alarms; the platform layer handles data storage, indicator calculation, and alarm aggregation processing; and the dashboard layer is responsible for visualizing operational status and presenting alarm information. The alarm mechanism, as the core means of anomaly early warning, typically configures corresponding rules at each of the three layers to adapt to the monitoring functions of each level.

[0003] However, the independent alarm rules and unidirectional data flow under the aforementioned layered architecture can easily lead to multi-dimensional collaborative breakdowns. First, the lack of a unified judgment standard among the three layers of alarm rules means that the same device anomaly may be repeatedly reported across multiple layers, or alarms may be missed due to misaligned rule thresholds, resulting in insufficient accuracy and reliability of alarms. Second, the indiscriminate push of massive amounts of alarm data to the dashboard via the platform can easily trigger an alarm storm, causing front-end rendering lag and page response delays, with valid anomaly information being overwhelmed by a large number of invalid alarms. Third, after an alarm event is triggered, the dashboard can only passively receive and display alarm information, unable to link back to the data collection end to retrieve high-frequency sampling data before and after the fault, making it difficult for maintenance personnel to quickly trace the root cause of the fault.

[0004] Current solutions to address these issues primarily focus on localized optimizations of single aspects, such as adding alarm clustering and noise reduction algorithms at the platform level or setting alarm hierarchical display rules at the dashboard. However, these solutions fail to break the inherent one-way flow of alarm data, fail to establish a two-way collaborative mechanism between the data acquisition end and the dashboard, and cannot fundamentally solve the systemic problems of alarm collaboration failure, dashboard load imbalance, and low fault tracing efficiency. Consequently, they are ill-suited to the real-time monitoring and maintenance needs of large-scale industrial equipment. Summary of the Invention

[0005] Based on this, the purpose of this invention is to provide a collaborative processing method and system for industrial equipment data acquisition and intelligent monitoring dashboards, in order to solve the inherent pattern of unidirectional flow of alarm data in the prior art, and the failure to establish a two-way collaborative mechanism between the acquisition end and the dashboard end, which leads to the inability to fundamentally solve the systemic problems of alarm collaboration failure, dashboard load imbalance and low fault tracing efficiency.

[0006] The first aspect of this invention proposes: A collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboards, wherein the method includes: The operation data of industrial equipment is processed by edge-level hierarchical encoding to generate a sampling encoding stream with multi-level sampling granularity. The rendering computing power pool is processed by sharding and quantization for each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile. Based on the partition computing power quota profile, the hierarchical coding contract of the collection terminal is generated in reverse. The abnormal pre-detection signal uploaded by the collection terminal is received synchronously and the computing power quota of the corresponding Kanban partition is reserved for it, so as to establish a collaborative contract link with bidirectional constraints between sampling coding and Kanban computing power. Independent alarm feature images are built at the acquisition end and the dashboard end respectively. The difference between the alarm feature images at both ends is verified through the collaborative contract link to generate a valid alarm set with consistent verification. The computing power partitioning boundary of each view partition of the dashboard is dynamically redrawn based on the valid alarm set. Based on the effective alarm set, the acquisition end is triggered to perform source tracing rollback encrypted sampling, generate source tracing encrypted data stream, and map the source tracing encrypted data stream to the source tracing dedicated view partition of the dashboard to form a source tracing link that links sampling granularity with view rendering.

[0007] The beneficial effects of this invention are as follows: This solution generates a sampling code stream with multi-level sampling granularity by performing edge-end hierarchical encoding processing on the industrial equipment operation data. Simultaneously, it performs rendering computing power pool fragmentation quantization on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile. Based on this profile, it reversely generates a hierarchical encoding contract at the acquisition end and reserves corresponding partition computing power quotas for anomaly pre-detection signals. It establishes a collaborative contract link with bidirectional constraints between sampling encoding and dashboard computing power, breaking the one-way flow mode of alarm data from the root and effectively solving the problem of alarm collaboration failure. It generates a valid alarm set by performing difference verification based on the dual-end independent alarm feature mirrors. Based on this, it dynamically redraws the computing power fragmentation boundary of the dashboard view partition, realizing dynamic allocation of computing power on demand and significantly improving the dashboard load imbalance. At the same time, it triggers the acquisition end to trace back and encrypt sampling based on the valid alarm set, generates a traceability encrypted data stream and maps it to the dashboard traceability dedicated view partition, forming a traceability link with sampling granularity and view rendering linkage, greatly improving the efficiency of fault traceability and systematically solving the core pain points in industrial monitoring scenarios.

[0008] Furthermore, the steps of performing edge-level hierarchical encoding processing on the operating data of industrial equipment to generate a sampling encoding stream containing multi-level sampling granularity, and performing rendering computing power pool fragmentation quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile include: The device operation data is hierarchically divided according to time resolution and feature dimension, generating three-level coded data including the original full data layer, feature aggregation layer, and trend summary layer, and an independent coded identifier is assigned to each layer of data. The rendering computing power, caching computing power and interaction computing power of each view partition of the dashboard are statistically analyzed in segments. The computing power baseline threshold and dynamic adjustable margin of each partition are calculated, and a partition computing power quota profile containing the baseline quota and adjustable margin is generated accordingly.

[0009] Furthermore, the step of generating a hierarchical coding contract for the collection end based on the partitioned computing power quota profile, synchronously receiving the anomaly pre-detection signal uploaded by the collection end and reserving the corresponding Kanban partition's computing power quota for it, in order to establish a collaborative contract link with bidirectional constraints between sampling coding and Kanban computing power, includes: Based on the baseline quota in the computing power quota profile of each partition, the upper limit of the encoding level and the data transmission rate of the corresponding acquisition channel are deduced in reverse, a hierarchical encoding contract is generated and sent to the edge acquisition unit. The receiver receives an anomaly pre-detection signal identified by the trend summary layer, allocates a preset proportion of computing power from the adjustable margin of the corresponding partition as reserved computing power, and updates the temporary coding permissions in the hierarchical coding contract. A dedicated transmission channel for binding coding contracts and computing power quotas is established between the acquisition terminal and the corresponding view partition of the dashboard, forming the collaborative contract link.

[0010] Furthermore, the step of constructing independent alarm feature mirrors at the acquisition end and the dashboard end respectively, and performing difference verification on the alarm feature mirrors at both ends through the collaborative contract link to generate a valid alarm set with consistent verification includes: The acquisition end extracts time-series fluctuation features based on the original full-volume data and constructs an alarm feature mirror on the acquisition side; the dashboard end extracts view aggregation features based on the received feature aggregation layer data and constructs an alarm feature mirror on the dashboard side. When an alarm candidate is triggered at either end, the alarm feature image of this end is synchronized to the other end through the collaborative contract link, and the feature dimension difference and temporal offset of the feature images of the two ends are calculated. When both the difference and the offset are within the preset allowable range, the alarm candidate is marked as a valid alarm and included in the valid alarm set.

[0011] Furthermore, the step of dynamically redrawing the computing power partition boundaries of each view partition of the dashboard based on the effective alarm set includes: Collect the view partitions and alarm levels corresponding to each alarm in the valid alarm set, and calculate the computing power increment required for each partition; According to a preset order, computing power is borrowed from the adjustable margin of non-alarm-related partitions and added to the alarm-related partitions. Based on the results of the computing power replenishment, the computing power sharding boundaries of each partition are redefined, the partition computing power quota profile is updated, and the updated quota information is synchronized to the collaborative contract link.

[0012] Furthermore, the step of triggering source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set to generate a source tracing encrypted data stream includes: Using the occurrence time of a valid alarm and the associated device topology as anchor points, determine the time window and device link range for source tracing and rollback; An encrypted sampling command is sent to the corresponding acquisition unit to raise the sampling granularity of the device link within the time window from the current level to the original full volume level, and to collect the missing fine-grained data of the rollback period. The fine-grained data collected is time-aligned and linked to generate the source-tracing encrypted data stream.

[0013] Furthermore, the step of mapping the source-tracing encrypted data stream to a source-tracing-specific view partition of the dashboard to form a source-tracing link that links sampling granularity with view rendering includes: Create a dedicated view partition for tracing the source in the dashboard and allocate an independent computing power quota to this partition; The source-tracing encrypted data stream is mapped to a source-specific view partition according to the device topology and time sequence, generating a time-series curve and topology link diagram with sampling granularity identifier; Establish a linkage between view zooming operations and sampling granularity. When the view zoom ratio changes, dynamically adjust the sampling encoding level for the corresponding time period through the collaborative contract link.

[0014] The second aspect of the present invention proposes: A collaborative processing system for industrial equipment data acquisition and intelligent monitoring dashboards, wherein the system includes: The quantization module is used to perform edge-level hierarchical encoding processing on the operating data of industrial equipment, generate a sampling encoding stream containing multi-level sampling granularity, and perform rendering computing power pool sharding quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile. The construction module is used to reverse generate the hierarchical coding contract of the collection terminal based on the partition computing power quota profile, synchronously receive the abnormal pre-detection signal uploaded by the collection terminal and reserve the computing power quota of the corresponding Kanban partition for it, so as to establish a collaborative contract link of bidirectional constraints between sampling coding and Kanban computing power. The verification module is used to build independent alarm feature images at the acquisition end and the dashboard end respectively, perform difference verification on the alarm feature images of the two ends through the collaborative contract link, generate a valid alarm set with consistent verification, and dynamically redraw the computing power partition boundaries of each view partition of the dashboard based on the valid alarm set. The linkage module is used to trigger the source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set, generate the source tracing encrypted data stream, and map the source tracing encrypted data stream to the source tracing dedicated view partition of the dashboard to form a source tracing link that links sampling granularity and view rendering.

[0015] Furthermore, the steps of performing edge-level hierarchical encoding processing on the operating data of industrial equipment to generate a sampling encoding stream containing multi-level sampling granularity, and performing rendering computing power pool fragmentation quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile include: The device operation data is hierarchically divided according to time resolution and feature dimension, generating three-level coded data including the original full data layer, feature aggregation layer, and trend summary layer, and an independent coded identifier is assigned to each layer of data. The rendering computing power, caching computing power and interaction computing power of each view partition of the dashboard are statistically analyzed in segments. The computing power baseline threshold and dynamic adjustable margin of each partition are calculated, and a partition computing power quota profile containing the baseline quota and adjustable margin is generated accordingly.

[0016] Furthermore, the step of generating a hierarchical coding contract for the collection end based on the partitioned computing power quota profile, synchronously receiving the anomaly pre-detection signal uploaded by the collection end and reserving the corresponding Kanban partition's computing power quota for it, in order to establish a collaborative contract link with bidirectional constraints between sampling coding and Kanban computing power, includes: Based on the baseline quota in the computing power quota profile of each partition, the upper limit of the encoding level and the data transmission rate of the corresponding acquisition channel are deduced in reverse, a hierarchical encoding contract is generated and sent to the edge acquisition unit. The receiver receives an anomaly pre-detection signal identified by the trend summary layer, allocates a preset proportion of computing power from the adjustable margin of the corresponding partition as reserved computing power, and updates the temporary coding permissions in the hierarchical coding contract. A dedicated transmission channel for binding coding contracts and computing power quotas is established between the acquisition terminal and the corresponding view partition of the dashboard, forming the collaborative contract link.

[0017] Furthermore, the step of constructing independent alarm feature mirrors at the acquisition end and the dashboard end respectively, and performing difference verification on the alarm feature mirrors at both ends through the collaborative contract link to generate a valid alarm set with consistent verification includes: The acquisition end extracts time-series fluctuation features based on the original full-volume data and constructs an alarm feature mirror on the acquisition side; the dashboard end extracts view aggregation features based on the received feature aggregation layer data and constructs an alarm feature mirror on the dashboard side. When an alarm candidate is triggered at either end, the alarm feature image of this end is synchronized to the other end through the collaborative contract link, and the feature dimension difference and temporal offset of the feature images of the two ends are calculated. When both the difference and the offset are within the preset allowable range, the alarm candidate is marked as a valid alarm and included in the valid alarm set.

[0018] Furthermore, the step of dynamically redrawing the computing power partition boundaries of each view partition of the dashboard based on the effective alarm set includes: Collect the view partitions and alarm levels corresponding to each alarm in the valid alarm set, and calculate the computing power increment required for each partition; According to a preset order, computing power is borrowed from the adjustable margin of non-alarm-related partitions and added to the alarm-related partitions. Based on the results of the computing power replenishment, the computing power sharding boundaries of each partition are redefined, the partition computing power quota profile is updated, and the updated quota information is synchronized to the collaborative contract link.

[0019] Furthermore, the step of triggering source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set to generate a source tracing encrypted data stream includes: Using the occurrence time of a valid alarm and the associated device topology as anchor points, determine the time window and device link range for source tracing and rollback; An encrypted sampling command is sent to the corresponding acquisition unit to raise the sampling granularity of the device link within the time window from the current level to the original full volume level, and to collect the missing fine-grained data of the rollback period. The fine-grained data collected is time-aligned and linked to generate the source-tracing encrypted data stream.

[0020] Furthermore, the step of mapping the source-tracing encrypted data stream to a source-tracing-specific view partition of the dashboard to form a source-tracing link that links sampling granularity with view rendering includes: Create a dedicated view partition for tracing the source in the dashboard and allocate an independent computing power quota to this partition; The source-tracing encrypted data stream is mapped to a source-specific view partition according to the device topology and time sequence, generating a time-series curve and topology link diagram with sampling granularity identifier; Establish a linkage between view zooming operations and sampling granularity. When the view zoom ratio changes, dynamically adjust the sampling encoding level for the corresponding time period through the collaborative contract link.

[0021] The third aspect of the present invention proposes: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboards as described above.

[0022] The fourth aspect of the present invention proposes: A computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboards as described above.

[0023] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0024] Figure 1 A flowchart illustrating the collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard provided in the first embodiment of the present invention; Figure 2 This is a structural block diagram of the collaborative processing system for industrial equipment data acquisition and intelligent monitoring dashboard provided in the third embodiment of the present invention.

[0025] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0026] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0027] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0029] Please see Figure 1 The image shows a collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard provided in the first embodiment of the present invention. The collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard provided in this embodiment can trigger the acquisition end to trace back and encrypt sampling based on the effective alarm set, generate traceability encrypted data stream and map it to the traceability-specific view partition of the dashboard, forming a traceability link with sampling granularity and view rendering linkage, which greatly improves the efficiency of fault traceability and systematically solves the core pain points in industrial monitoring scenarios.

[0030] Specifically, this embodiment provides: A collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboards, wherein the method includes: Step S10: Perform edge-end layered encoding processing on the operating data of industrial equipment to generate a sampling encoding stream containing multi-level sampling granularity, and perform rendering computing power pool fragmentation quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile. It's important to note that edge-level hierarchical encoding processing is performed on the operational data of industrial equipment to generate a sampling encoding stream with multi-level sampling granularity. This gives the collected data the flexibility of "adjustable granularity and on-demand retrieval," allowing for low-granularity data to save bandwidth under normal conditions and rapid granularity escalation during anomalies. Simultaneously, rendering computing power pool fragmentation and quantification processing is performed on each view partition of the intelligent monitoring dashboard to generate partition computing power quota profiles. This transforms the dashboard's computing power from an uncontrollable overall resource pool into manageable, adjustable, and releasable partitioned elastic resources. This step is the foundation of the entire collaborative system, establishing quantifiable and matchable resource measurement standards for both the data acquisition end and the dashboard end, providing accurate data basis for subsequent two-way constraints. Taking the stamping workshop as an example, the edge generates a three-level encoding stream for the vibration, temperature, current, and pressure data of each press, while the dashboard end divides the interface into four major partitions: production line overview, equipment details, energy consumption statistics, and output statistics, and completes computing power quantification. Both ends achieve standardized measurement of resources.

[0031] Step S20: Based on the partition computing power quota profile, generate the hierarchical coding contract of the collection terminal in reverse, synchronously receive the abnormal pre-detection signal uploaded by the collection terminal and reserve the computing power quota of the corresponding Kanban partition for it, so as to establish a collaborative contract link of bidirectional constraint between sampling coding and Kanban computing power. It's worth noting that the hierarchical coding contract for the acquisition end is generated in reverse based on the partitioned computing power quota profile of the dashboard. The dashboard's computing power capacity is used to constrain the coding granularity and transmission rate of the acquisition end, fundamentally preventing screen stuttering or resource waste caused by mismatches between data upload and rendering capabilities. Simultaneously, it receives anomaly pre-detection signals uploaded by the acquisition end, reserving computing power quotas for the corresponding dashboard partitions in advance to ensure resource response speed in abnormal scenarios. This process breaks the traditional one-way push model of "what the acquisition end sends, what the dashboard receives," establishing a two-way constraint mechanism of "computing power constraining the coding upper limit, and coding adapting to computing power supply," achieving dynamic matching of data traffic and rendering computing power.

[0032] Step S30: Construct independent alarm feature images at the acquisition end and the dashboard end respectively. Perform difference verification on the alarm feature images of the two ends through the collaborative contract link to generate a valid alarm set with consistent verification. Dynamically redraw the computing power partition boundaries of each view partition of the dashboard based on the valid alarm set. It's worth noting that independent alarm feature mirrors are constructed at both the data acquisition and dashboard ends. Leveraging the complementarity of data granularity and feature dimensions at both ends, a cross-validation of differences is performed through a collaborative contract link. This filters out false alarms caused by noise and interference from one end, generating a valid alarm set with consistent validation. Simultaneously, based on the level and spatial distribution of valid alarms, the computing power partition boundaries of each view area on the dashboard are dynamically redefined, directing computing resources towards higher-priority alarm-related partitions. This process significantly improves alarm accuracy through dual-end validation and maximizes resource efficiency through dynamic scheduling of existing computing power, ensuring rendering performance in abnormal scenarios without requiring hardware expansion.

[0033] Step S40: Based on the valid alarm set, trigger the source tracing rollback encrypted sampling at the acquisition end, generate the source tracing encrypted data stream, and map the source tracing encrypted data stream to the source tracing dedicated view partition of the dashboard to form a source tracing link that links sampling granularity with view rendering.

[0034] It's worth noting that, based on the valid alarm set, the data acquisition end triggers backtracking and encrypted sampling, replenishing the full set of fine-grained data for the alarm period as needed, generating a traceability encrypted data stream with topological tags. This high-precision traceability data is then mapped to a dedicated traceability view partition on the dashboard. Simultaneously, a linkage between view zooming interaction and sampling granularity is established, forming a traceability link where the sampling granularity dynamically adjusts according to view requirements. This process achieves end-to-end linkage between "alarm confirmation—data replenishment—visual analysis," making fault traceability both efficient and accurate, completely resolving the pain points of slow traceability and data / view disconnect in traditional architectures.

[0035] Second Embodiment Furthermore, the steps of performing edge-level hierarchical encoding processing on the operating data of industrial equipment to generate a sampling encoding stream containing multi-level sampling granularity, and performing rendering computing power pool fragmentation quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile include: The device operation data is hierarchically divided according to time resolution and feature dimension, generating three-level coded data including the original full data layer, feature aggregation layer, and trend summary layer, and an independent coded identifier is assigned to each layer of data. The rendering computing power, caching computing power and interaction computing power of each view partition of the dashboard are statistically analyzed in segments. The computing power baseline threshold and dynamic adjustable margin of each partition are calculated, and a partition computing power quota profile containing the baseline quota and adjustable margin is generated accordingly.

[0036] It's important to note that the device operation data is hierarchically divided based on two dimensions: time resolution and feature dimension. This results in three levels of coded data: the raw full data layer, the feature aggregation layer, and the trend summary layer. Each layer is assigned a unique coded identifier. The raw full data layer contains the highest sampling frequency of raw sensor data, offering the highest accuracy and largest data volume. It is stored locally on the edge gateway by default and is not actively uploaded to the dashboard. The feature aggregation layer contains minute-level statistical feature data, including core indicators such as mean, peak value, volatility, and kurtosis. With a moderate data volume, it is the primary transmission layer for routine monitoring. The trend summary layer contains coarse-grained operation status labels and trend slopes, with the smallest data volume, used for global overview display and lightweight edge anomaly detection. This three-level layering design allows for multi-granularity flexibility in data collection. Under normal conditions, low-granularity data saves bandwidth, while in case of anomalies, high-granularity data can be retrieved as needed, balancing resource consumption and data accuracy. Taking the No. 3 servo press as an example, the original full data layer contains 48 channels of original sensor data, including slider vibration, machine body temperature, main motor current, and stamping pressure, sampled at 10Hz. The local storage period is 7 days, and the daily data volume of a single device is about 12GB. The feature aggregation layer contains statistical features at the minute level, including the mean, peak value, root mean square, kurtosis, and number of times the limit is exceeded for each parameter. The daily data volume of a single device is about 80MB, which is the normal upload level. The trend summary layer contains the running status labels and trend slope at the 5-minute level. The data volume is only 1 / 20 of that of the feature aggregation layer and is used for global overview and rapid edge anomaly prediction.

[0037] For each view partition of the intelligent monitoring dashboard, statistics are calculated from three dimensions: rendering computing power, caching computing power, and interaction computing power. Rendering computing power corresponds to the drawing overhead of charts, animations, and 3D models of equipment within the partition; caching computing power corresponds to the memory overhead of historical data preloading and real-time data caching; and interaction computing power corresponds to the real-time response overhead of user zooming, dragging, and pop-up operations. Based on this, a baseline threshold and dynamic adjustable margin for computing power of each partition are calculated. The baseline threshold is used to ensure the normal and stable operation of the partition, while the dynamic adjustable margin serves as an elastic resource pool for scheduling in abnormal scenarios. Finally, a partition computing power quota profile containing the baseline quota and adjustable margin is generated. The profile fully quantifies the computing power base and elastic space of each view partition, which is the core basis for subsequent reverse derivation of coding contracts and cross-partition computing power sharing. Taking the stamping workshop monitoring dashboard as an example, the dashboard is divided into four view partitions: production line overview partition, single equipment details partition, energy consumption statistics partition, and production statistics partition. According to the segmented statistics, the baseline computing power quota for the single device details partition is 1.2 TOPS, used to support 6-channel real-time curves, 3D device model rendering, and basic interaction, with a dynamically adjustable margin of 0.5 TOPS; the baseline quota for the production line overview partition is 0.8 TOPS, with an adjustable margin of 0.2 TOPS; and the baseline quota for energy consumption and production statistics partitions is 0.5 TOPS each, with an adjustable margin of 0.3 TOPS each. The computing power boundaries of each partition are clear, and the flexibility is well-defined, providing a precise quantitative benchmark for subsequent two-way constraints.

[0038] Furthermore, the step of generating a hierarchical coding contract for the collection end based on the partitioned computing power quota profile, synchronously receiving the anomaly pre-detection signal uploaded by the collection end and reserving the corresponding Kanban partition's computing power quota for it, in order to establish a collaborative contract link with bidirectional constraints between sampling coding and Kanban computing power, includes: Based on the baseline quota in the computing power quota profile of each partition, the upper limit of the encoding level and the data transmission rate of the corresponding acquisition channel are deduced in reverse, a hierarchical encoding contract is generated and sent to the edge acquisition unit. The receiver receives an anomaly pre-detection signal identified by the trend summary layer, allocates a preset proportion of computing power from the adjustable margin of the corresponding partition as reserved computing power, and updates the temporary coding permissions in the hierarchical coding contract. A dedicated transmission channel for binding coding contracts and computing power quotas is established between the acquisition terminal and the corresponding view partition of the dashboard, forming the collaborative contract link.

[0039] It's important to note that, based on the baseline quota in the computing power quota profile of each partition, the highest encoding level and maximum data transmission rate that the corresponding acquisition channel can support are derived in reverse. This generates a hierarchical encoding contract, which is then distributed to the edge acquisition units. Partitions with high computing power correspond to higher encoding level permissions and higher transmission rate limits, while overview partitions with low computing power correspond to lower encoding levels. This fundamentally avoids the problems of dashboard rendering lag and computing power overload caused by blindly uploading high-bitrate data at the acquisition end in traditional architectures. It also avoids resource waste caused by sufficient computing power but excessively low data granularity, achieving a precise match between the amount of collected data and the dashboard computing power. For example, the production line overview partition has limited computing power, and its corresponding encoding contract only allows the transmission of trend summary layer data, with a single device transmission rate limit of 0.5Mbps, preventing a large influx of data from causing overview screen lag. The single device details partition has sufficient computing power, and its corresponding contract normally allows the transmission of feature aggregation layer data, with a transmission rate limit of 5Mbps, meeting the accuracy requirements of regular monitoring. After the contract is issued to the edge data acquisition gateway, the gateway strictly uploads data according to the agreed level. Under normal circumstances, the bandwidth usage of the entire workshop is reduced from 82Mbps in the traditional solution to 31Mbps, and the bandwidth resource utilization is greatly optimized.

[0040] Upon receiving an anomaly pre-detection signal from the acquisition end based on the trend summary layer, even though the alarm has not yet been definitively confirmed, a clear anomaly risk already exists. The system immediately allocates a preset proportion of computing power from the adjustable margin of the corresponding view partition as reserved computing power. Simultaneously, it updates the temporary coding permissions in the hierarchical coding contract, temporarily allowing the acquisition end to upload more granular feature aggregation layer full-dimensional data. This pre-reservation mechanism enables advance resource preparation for anomaly risks, avoiding response delays caused by requesting computing power after the alarm is officially triggered, and ensuring the speed of resource supply in anomaly scenarios. For example, when the edge gateway identifies a continuously rising abnormal sign of vibration parameters in servo press No. 3 based on trend summary layer data, it immediately uploads an anomaly pre-detection signal to the dashboard. After receiving the signal, the dashboard immediately allocates 30% (0.15 TOPS) of the 0.5 TOPS adjustable margin from the corresponding detail partition of the No. 3 servo pressure machine as reserved computing power. At the same time, it updates the temporary permissions of the hierarchical coding contract, allowing the edge to temporarily upgrade to the full-dimensional feature aggregation layer for transmission. This prepares resources in advance for possible subsequent alarm confirmation. Compared with scheduling computing power after the alarm is triggered, the response speed is improved by about 60%.

[0041] A dedicated transmission channel is established between the acquisition unit at the acquisition end and the corresponding view partition at the dashboard end, simultaneously binding the coding contract rules and computing power quotas. Data transmission, command interaction, and contract updates within this channel are independent of ordinary monitoring links, possessing higher transmission priority and resource guarantees. This dedicated channel deeply binds the coding rules and computing power quotas, forming a stable collaborative contract link. Subsequent alarm feature synchronization, source data transmission, and computing power adjustment commands are all completed through this link, ensuring the determinism and timeliness of collaborative interactions. For example, a dedicated TCP transmission channel is established between the No. 3 servo pressure machine and its corresponding details partition, set with the highest QoS priority. The channel simultaneously binds the currently effective coding contract and computing power quota. All anomaly-related feature data and control commands are transmitted through this channel, unaffected by traffic fluctuations of other ordinary monitoring data, with transmission latency consistently controlled within 10ms, ensuring the real-time nature of collaborative commands.

[0042] Furthermore, the step of constructing independent alarm feature mirrors at the acquisition end and the dashboard end respectively, and performing difference verification on the alarm feature mirrors at both ends through the collaborative contract link to generate a valid alarm set with consistent verification includes: The acquisition end extracts time-series fluctuation features based on the original full-volume data and constructs an alarm feature mirror on the acquisition side; the dashboard end extracts view aggregation features based on the received feature aggregation layer data and constructs an alarm feature mirror on the dashboard side. When an alarm candidate is triggered at either end, the alarm feature image of this end is synchronized to the other end through the collaborative contract link, and the feature dimension difference and temporal offset of the feature images of the two ends are calculated. When both the difference and the offset are within the preset allowable range, the alarm candidate is marked as a valid alarm and included in the valid alarm set.

[0043] It should be noted that the acquisition end extracts high-precision time-series fluctuation features from the locally stored raw full-layer data to construct an alarm feature mirror on the acquisition side. This mirror data has high precision and rich time-domain details, but its feature dimensions are limited by edge computing power, focusing on the identification of time-domain fluctuation anomalies of single devices. The dashboard end extracts multi-dimensional view aggregation features from the received feature aggregation layer data to construct an alarm feature mirror on the dashboard side. This mirror data has slightly coarser granularity, but it covers broad-dimensional features such as device association, multi-parameter linkage, and process matching, focusing on global correlation anomalies and process logic verification. The independent construction of mirrors at both ends avoids the bandwidth pressure of cross-end transmission of full data and utilizes the complementarity of the data at both ends to form a dual verification basis, achieving high reliability verification with low overhead. Taking the No. 3 servo press as an example, the alarm feature mirror on the acquisition side is constructed based on local 10Hz raw vibration data, extracting 24-dimensional refined time-frequency features such as time-domain peak value, frequency-domain characteristic frequency, number of impact pulses, and kurtosis coefficient, focusing on identifying fine waveform features of mechanical anomalies such as bearing wear and gear failure. The alarm feature mirror on the dashboard side is constructed based on received minute-level aggregated data, extracting 18-dimensional aggregated features such as the linkage correlation of vibration, temperature, and current, parameter transmission characteristics of upstream and downstream equipment, and parameter dispersion of equipment in the same batch, focusing on judging the rationality of anomalies from the global process logic. The two ends are modeled independently and complement each other, transmitting only hundred-byte-level feature vectors, and the bandwidth overhead is negligible.

[0044] When either end triggers an alarm candidate, the alarm feature mirror of this end is synchronized to the other end via the collaborative contract link. Two core metrics are calculated: feature dimension difference and timing offset. Feature dimension difference determines whether the abnormal features identified by both ends belong to the same fault mode; a higher difference indicates a greater deviation in the abnormal features identified by both ends. Timing offset determines whether the occurrence time of the alarms at both ends is within the allowable clock deviation range, eliminating false matches caused by timing misalignment. Dual-dimensional verification effectively filters out single-end misjudgments. For example, a false alarm triggered by a momentary sensor jump at the edge end, which has no corresponding associated anomaly in the aggregated features on the dashboard side, will be filtered out due to excessive difference. For instance, when the edge end of servo press No. 3 detects a momentary exceedance of vibration and triggers an alarm candidate, the 24-dimensional acquisition side feature mirror is synchronized to the dashboard side via the collaborative contract link. The dashboard compares the observed feature with its own aggregated feature mirror image. The calculated feature dimension difference is 4.2% and the time offset is 1.8 seconds, both within the preset tolerance range of 10% difference and 5-second offset, indicating that the two ends have a high degree of consistency in judging anomalies. However, if the edge device triggers an alarm candidate due to a single sensor jump caused by electromagnetic interference in the workshop, and there is no corresponding temperature or current linkage anomaly in the aggregated features on the dashboard side, the feature difference will reach 42%, which will be judged as a false alarm and filtered out.

[0045] When both the feature dimension difference and the time series offset are within the preset allowable range, it indicates that both ends agree on the anomaly, confirming a genuine equipment failure. The alarm candidate is then marked as a valid alarm and included in the valid alarm set. The dual-end cross-validation mechanism eliminates the need to transmit all original data; validation can be completed using only lightweight feature mirroring. With almost no increase in bandwidth overhead, it significantly reduces false alarms caused by electromagnetic interference, sensor noise, and data packet loss in industrial settings, thus significantly improving the reliability of the monitoring system. Actual testing in the stamping workshop showed that after adopting the dual-end validation mechanism, the overall false alarm rate decreased from 27.6% in the traditional single-end solution to 5.8%, and the number of invalid alarms decreased by nearly 80%, greatly reducing the interference of invalid alarms on maintenance personnel and improving maintenance efficiency.

[0046] Furthermore, the step of dynamically redrawing the computing power partition boundaries of each view partition of the dashboard based on the effective alarm set includes: Collect the view partitions and alarm levels corresponding to each alarm in the valid alarm set, and calculate the computing power increment required for each partition; According to a preset order, computing power is borrowed from the adjustable margin of non-alarm-related partitions and added to the alarm-related partitions. Based on the results of the computing power replenishment, the computing power sharding boundaries of each partition are redefined, the partition computing power quota profile is updated, and the updated quota information is synchronized to the collaborative contract link.

[0047] It's important to note that the statistical analysis of valid alarms includes the view partition and alarm level corresponding to each alarm. Based on the rendering requirements of each alarm level, including data refresh frequency, number of linked charts, fault animation simulation, and topology highlighting effects, the incremental computing power required for each alarm-related partition is calculated. Higher alarm levels require more rendering and interaction computing power, resulting in a larger incremental computing power, ensuring that computing power allocation is strongly tied to business priorities and that core faults receive sufficient resource guarantees. For example, the stamping workshop divides alarms into three levels: Level 1 alarms are serious faults such as equipment shutdown, requiring an additional 0.3 TOPS of computing power to support high-frequency data refresh, 3D fault animation simulation, and 8-way linked parameter display; Level 2 alarms are anomalies such as parameter exceeding limits, requiring an additional 0.15 TOPS of computing power to support curve refinement and related parameter display; Level 3 alarms are minor warnings, requiring no additional computing power. In this example, the No. 3 servo press triggers a Level 1 alarm, requiring an additional 0.3 TOPS of computing power in the corresponding details partition to ensure rendering performance for fault analysis.

[0048] The system prioritizes view partitions from lowest to highest business priority, borrowing computing power from the adjustable margin of low-priority partitions not associated with alarms to supplement high-priority partitions associated with alarms. This mechanism does not require expanding total computing power resources; it ensures smooth rendering of alarm scenarios through dynamic scheduling of existing resources, maximizing the efficiency of limited computing power. For example, in the stamping workshop, the view priorities from highest to lowest are: equipment details partition, production line overview partition, energy consumption statistics partition, and production statistics partition. If the alarm on servo press No. 3 requires an additional 0.3 TOPS of computing power, the system first borrows 0.2 TOPS of adjustable margin from the lowest priority production statistics partition, and then borrows 0.1 TOPS of adjustable margin from the next lowest priority energy consumption statistics partition, totaling 0.3 TOPS, which is fully supplemented to the alarm details partition. The borrowing process is strictly limited to the adjustable margin range, without touching the baseline computing power quota of each partition, ensuring that the basic monitoring functions of low-priority partitions are not affected.

[0049] Based on the results of the computing power sharing, the computing power sharding boundaries of each partition are redefined, the baseline quota and adjustable margin values ​​in the partition computing power quota profile are updated, and the updated quota information is synchronized to the collaborative contract link, correspondingly adjusting the hierarchical coding permissions of the acquisition end. After the computing power redistribution, the alarm partition receives more computing power, which can support fine-grained data with higher bitrates, and the corresponding acquisition coding permissions are simultaneously upgraded; the computing power of the partition that was shared is reduced, and the corresponding coding permissions are simultaneously downgraded, always maintaining a two-way match between computing power and coding. This forms a complete closed loop of "alarm triggering - computing power redistribution - contract update - coding adaptation". In this example, after the computing power adjustment, the baseline quota of the No. 3 servo pressure machine details partition is temporarily increased to 1.5 TOPS, and the adjustable margin remains at 0.35 TOPS; the adjustable margin of the production partition is reduced to 0.1 TOPS, and the adjustable margin of the energy consumption partition is reduced to 0.2 TOPS. The updated quota information is synchronized to the collaborative contract link, and the coding permissions of the No. 3 servo pressure machine are upgraded accordingly. Temporary support for fine-grained data transmission up to 12Mbps is provided to ensure that the data accuracy of alarm analysis and the smoothness of rendering are improved simultaneously.

[0050] Furthermore, the step of triggering source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set to generate a source tracing encrypted data stream includes: Using the occurrence time of a valid alarm and the associated device topology as anchor points, determine the time window and device link range for source tracing and rollback; An encrypted sampling command is sent to the corresponding acquisition unit to raise the sampling granularity of the device link within the time window from the current level to the original full volume level, and to collect the missing fine-grained data of the rollback period. The fine-grained data collected is time-aligned and linked to generate the source-tracing encrypted data stream.

[0051] It's important to note that the time frame for tracing the source is determined by using the occurrence of a valid alarm as the time anchor point and then moving backward by a preset duration to cover the symptom-developing stage before the fault occurs. Simultaneously, starting from the alarming device, the process topology of the industrial equipment is extended upstream and downstream to determine the scope of the equipment link requiring tracing, covering upstream and downstream related equipment that may lead to the fault. This dual-dimensional anchoring of time and topology avoids data redundancy caused by full-time, full-equipment supplementary data collection while ensuring the data coverage required for root cause analysis, achieving precise delineation of the source tracing data. For example, if the alarm of servo press No. 3 occurs at 14:32:15, the system uses this time as an anchor point to go back 30 minutes to determine the source tracing time window as 14:02:00 to 14:32:15, fully covering the entire process of fault development from the appearance of symptoms to the final trigger. At the same time, with servo press No. 3 as the core node, the system extends upstream along the process topology to the sheet metal feeder and destacking robot, and downstream to the discharge conveyor belt and online inspection station, covering a total of 8 related devices, fully covering the possible transmission links of the fault, and ensuring that the root cause analysis is thorough.

[0052] The system issues encrypted sampling commands to all acquisition units within the traceability scope, raising the sampling granularity within the traceability time window from the current normal level to the original full-volume level. It also retrieves historical data stored locally on the edge to supplement fine-grained data that was not uploaded during the rollback period. This rollback and supplementary sampling mode fully utilizes the local storage capacity of the edge, transmitting only low-granularity data under normal conditions to save bandwidth, and then supplementing the full-volume data as needed after fault confirmation, perfectly balancing daily bandwidth consumption and the accuracy requirements of fault traceability. Simultaneously, the supplementary data is encrypted end-to-end throughout the process, ensuring the secure transmission of core operating parameters of industrial equipment. In this example, the system issues encrypted sampling commands to the acquisition gateways of 8 associated devices, retrieves historical data stored locally on the gateways, supplements the 10Hz full-volume raw data that was not uploaded within 30 minutes, and uses the AES-256 algorithm to encrypt the supplementary data end-to-end, ensuring the secure transmission of core process parameters. The entire rollback and supplementary sampling process takes only 28 seconds, far faster than the 20 minutes required for manual data retrieval from the historical database in traditional solutions, significantly improving the fault traceability response speed.

[0053] The system performs unified time-series alignment on the supplementary multi-device fine-grained data, eliminating clock deviations between different acquisition units based on the workshop NTP clock system to ensure high consistency of the time axis of multi-device data. Each data point is also labeled with its corresponding device link topology tag, clearly identifying the device node and link location, ultimately generating a structured, encrypted data stream for traceability. Time-series alignment and topology tagging provide a standardized data foundation for subsequent visualization on the dashboard, ensuring that the traceability view accurately reconstructs the temporal logic of fault evolution and device relationships. In this example, after supplementary data collection, the system performs unified time-series alignment on over 400 data points from 8 devices, controlling cross-device clock deviations to within 10ms, and labels each data point with a corresponding topology node, generating a structured, encrypted data stream for traceability. This provides a high-precision data foundation with consistent timing and clear topology for subsequent dashboard visualization analysis.

[0054] Furthermore, the step of mapping the source-tracing encrypted data stream to a source-tracing-specific view partition of the dashboard to form a source-tracing link that links sampling granularity with view rendering includes: Create a dedicated view partition for tracing the source in the dashboard and allocate an independent computing power quota to this partition; The source-tracing encrypted data stream is mapped to a source-specific view partition according to the device topology and time sequence, generating a time-series curve and topology link diagram with sampling granularity identifier; Establish a linkage between view zooming operations and sampling granularity. When the view zoom ratio changes, dynamically adjust the sampling encoding level for the corresponding time period through the collaborative contract link.

[0055] It's important to note that a dedicated traceability view partition is created within the intelligent monitoring dashboard, and this partition is allocated a separate computing power quota. This computing power is drawn from the globally adjustable reserve and does not consume the baseline quota of the regular monitoring partitions. This independent partition and independent computing power design achieves resource isolation between regular monitoring and fault traceability operations. Users conducting traceability analysis will not consume rendering resources from the regular monitoring screens, ensuring that production monitoring and fault analysis do not interfere with each other. For example, a dedicated traceability view partition occupying 30% of the dashboard's space is created on the right side, with 0.8 TOPS of independent computing power allocated from the global computing power pool. This does not consume the baseline resources of the four regular partitions. When maintenance personnel perform fault traceability analysis, the production line overview and real-time monitoring screens on the main monitoring interface remain smooth and will not experience lag due to traceability calculations.

[0056] The encrypted data stream for source tracing is mapped to a dedicated source tracing view partition according to the device topology and time sequence, generating high-precision time-series curves and device topology link diagrams with sampling granularity labels. The time-series curves visually display the fine fluctuations of each parameter before the fault, the topology link diagram clearly shows the fault propagation relationship between upstream and downstream devices, and the sampling granularity labels allow users to clearly understand the accuracy level of the current data. This visualization mapping transforms structured source tracing data into an intuitive analysis interface, supporting maintenance personnel in quickly locating the root cause of the fault. In this example, the source tracing partition simultaneously presents two core views: first, a multi-parameter overlay time-series curve, showing the fine changes in vibration, temperature, current, and pressure within 30 minutes before and after the fault, with a granularity label next to the curve indicating "10ms raw full data"; second, a device topology link heatmap, using color depth to indicate the degree of anomaly at each node, and arrows indicating the direction of fault propagation. Maintenance personnel can visually see the complete process of the fault starting from the feeder offset, gradually propagating to the servo press, and ultimately causing excessive vibration, improving root cause location efficiency by more than 60%.

[0057] A linkage rule is established between view zooming and sampling granularity. When a user zooms in to view local time details, the system automatically requests an increase in the sampling encoding level for the corresponding time period from the acquisition end via a collaborative contract link, retrieving data with finer granularity. When a user zooms out to view the overall trend over a longer period, the encoding level is automatically reduced, decreasing data transmission and rendering computational power consumption. This linkage mechanism achieves precise adaptation to "what you see is what you need." Data accuracy is automatically increased when users focus on details, and resources are automatically saved when browsing the overall picture. The sampling granularity is dynamically adjusted according to the user's analysis needs, forming an intelligent traceability link with bidirectional linkage between sampling and rendering. For example, when a user selects and zooms in to a 10-second local time window to view the detailed features of the vibration waveform, the system automatically requests higher-precision raw waveform data for that time period from the edge via a collaborative contract link, and the sampling encoding level is simultaneously increased, refining the curve to a single-cycle vibration waveform in real time. When the user zooms out to a 2-hour long-term trend, the system automatically lowers the encoding level to the feature aggregation layer, reducing unnecessary data transmission and rendering computational power consumption. This linkage mechanism ensures the accuracy of source tracing analysis while maximizing the saving of bandwidth and computing resources, achieving the optimal balance between user experience and resource efficiency.

[0058] Please see Figure 2 The third embodiment of the present invention provides: A collaborative processing system for industrial equipment data acquisition and intelligent monitoring dashboards, wherein the system includes: The quantization module is used to perform edge-level hierarchical encoding processing on the operating data of industrial equipment, generate a sampling encoding stream containing multi-level sampling granularity, and perform rendering computing power pool sharding quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile. The construction module is used to reverse generate the hierarchical coding contract of the collection terminal based on the partition computing power quota profile, synchronously receive the abnormal pre-detection signal uploaded by the collection terminal and reserve the computing power quota of the corresponding Kanban partition for it, so as to establish a collaborative contract link of bidirectional constraints between sampling coding and Kanban computing power. The verification module is used to build independent alarm feature images at the acquisition end and the dashboard end respectively, perform difference verification on the alarm feature images of the two ends through the collaborative contract link, generate a valid alarm set with consistent verification, and dynamically redraw the computing power partition boundaries of each view partition of the dashboard based on the valid alarm set. The linkage module is used to trigger the source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set, generate the source tracing encrypted data stream, and map the source tracing encrypted data stream to the source tracing dedicated view partition of the dashboard to form a source tracing link that links sampling granularity and view rendering.

[0059] Furthermore, the steps of performing edge-level hierarchical encoding processing on the operating data of industrial equipment to generate a sampling encoding stream containing multi-level sampling granularity, and performing rendering computing power pool fragmentation quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile include: The device operation data is hierarchically divided according to time resolution and feature dimension, generating three-level coded data including the original full data layer, feature aggregation layer, and trend summary layer, and an independent coded identifier is assigned to each layer of data. The rendering computing power, caching computing power and interaction computing power of each view partition of the dashboard are statistically analyzed in segments. The computing power baseline threshold and dynamic adjustable margin of each partition are calculated, and a partition computing power quota profile containing the baseline quota and adjustable margin is generated accordingly.

[0060] Furthermore, the step of generating a hierarchical coding contract for the collection end based on the partitioned computing power quota profile, synchronously receiving the anomaly pre-detection signal uploaded by the collection end and reserving the corresponding Kanban partition's computing power quota for it, in order to establish a collaborative contract link with bidirectional constraints between sampling coding and Kanban computing power, includes: Based on the baseline quota in the computing power quota profile of each partition, the upper limit of the encoding level and the data transmission rate of the corresponding acquisition channel are deduced in reverse, a hierarchical encoding contract is generated and sent to the edge acquisition unit. The receiver receives an anomaly pre-detection signal identified by the trend summary layer, allocates a preset proportion of computing power from the adjustable margin of the corresponding partition as reserved computing power, and updates the temporary coding permissions in the hierarchical coding contract. A dedicated transmission channel for binding coding contracts and computing power quotas is established between the acquisition terminal and the corresponding view partition of the dashboard, forming the collaborative contract link.

[0061] Furthermore, the step of constructing independent alarm feature mirrors at the acquisition end and the dashboard end respectively, and performing difference verification on the alarm feature mirrors at both ends through the collaborative contract link to generate a valid alarm set with consistent verification includes: The acquisition end extracts time-series fluctuation features based on the original full-volume data and constructs an alarm feature mirror on the acquisition side; the dashboard end extracts view aggregation features based on the received feature aggregation layer data and constructs an alarm feature mirror on the dashboard side. When an alarm candidate is triggered at either end, the alarm feature image of this end is synchronized to the other end through the collaborative contract link, and the feature dimension difference and temporal offset of the feature images of the two ends are calculated. When both the difference and the offset are within the preset allowable range, the alarm candidate is marked as a valid alarm and included in the valid alarm set.

[0062] Furthermore, the step of dynamically redrawing the computing power partition boundaries of each view partition of the dashboard based on the effective alarm set includes: Collect the view partitions and alarm levels corresponding to each alarm in the valid alarm set, and calculate the computing power increment required for each partition; According to a preset order, computing power is borrowed from the adjustable margin of non-alarm-related partitions and added to the alarm-related partitions. Based on the results of the computing power replenishment, the computing power sharding boundaries of each partition are redefined, the partition computing power quota profile is updated, and the updated quota information is synchronized to the collaborative contract link.

[0063] Furthermore, the step of triggering source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set to generate a source tracing encrypted data stream includes: Using the occurrence time of a valid alarm and the associated device topology as anchor points, determine the time window and device link range for source tracing and rollback; An encrypted sampling command is sent to the corresponding acquisition unit to raise the sampling granularity of the device link within the time window from the current level to the original full volume level, and to collect the missing fine-grained data of the rollback period. The fine-grained data collected is time-aligned and linked to generate the source-tracing encrypted data stream.

[0064] Furthermore, the step of mapping the source-tracing encrypted data stream to a source-tracing-specific view partition of the dashboard to form a source-tracing link that links sampling granularity with view rendering includes: Create a dedicated view partition for tracing the source in the dashboard and allocate an independent computing power quota to this partition; The source-tracing encrypted data stream is mapped to a source-specific view partition according to the device topology and time sequence, generating a time-series curve and topology link diagram with sampling granularity identifier; Establish a linkage between view zooming operations and sampling granularity. When the view zoom ratio changes, dynamically adjust the sampling encoding level for the corresponding time period through the collaborative contract link.

[0065] The fourth embodiment of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard as described above.

[0066] The fifth embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, it implements the collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard as described above.

[0067] In summary, the collaborative processing method and system for industrial equipment data acquisition and intelligent monitoring dashboards provided in the above embodiments of the present invention can trigger the acquisition end to trace back and encrypt sampling based on an effective alarm set, generate traceability encrypted data streams and map them to the dashboard traceability-specific view partitions, forming a traceability link that links sampling granularity with view rendering, greatly improving fault traceability efficiency and systematically solving the core pain points in industrial monitoring scenarios.

[0068] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0069] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0070] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0071] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0072] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0073] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboards, characterized in that, The method includes: The operation data of industrial equipment is processed by edge-level hierarchical encoding to generate a sampling encoding stream with multi-level sampling granularity. The rendering computing power pool is processed by sharding and quantization to generate a partition computing power quota profile for each view partition of the intelligent monitoring dashboard. Based on the partition computing power quota profile, the hierarchical coding contract of the collection terminal is generated in reverse. The abnormal pre-detection signal uploaded by the collection terminal is received synchronously and the computing power quota of the corresponding Kanban partition is reserved for it, so as to establish a collaborative contract link with bidirectional constraints between sampling coding and Kanban computing power. Independent alarm feature images are constructed at the acquisition end and the dashboard end respectively. The difference between the alarm feature images at both ends is verified through the collaborative contract link to generate a valid alarm set with consistent verification. The computing power partitioning boundary of each view partition of the dashboard is dynamically redrawn based on the valid alarm set. Based on the effective alarm set, the acquisition end is triggered to perform source tracing rollback encrypted sampling, generate source tracing encrypted data stream, and map the source tracing encrypted data stream to the source tracing dedicated view partition of the dashboard to form a source tracing link that links sampling granularity with view rendering.

2. The collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard according to claim 1, characterized in that, The steps of performing edge-level hierarchical encoding processing on the operating data of industrial equipment to generate a sampling encoding stream containing multi-level sampling granularity, and performing rendering computing power pool fragmentation quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile include: The device operation data is hierarchically divided according to time resolution and feature dimension, generating three-level coded data including the original full data layer, feature aggregation layer, and trend summary layer, and an independent coded identifier is assigned to each layer of data. The rendering computing power, caching computing power and interaction computing power of each view partition of the dashboard are statistically analyzed in segments. The computing power baseline threshold and dynamic adjustable margin of each partition are calculated, and a partition computing power quota profile containing the baseline quota and adjustable margin is generated accordingly.

3. The collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard according to claim 2, characterized in that, The steps of generating a hierarchical coding contract for the collection end based on the partitioned computing power quota profile, synchronously receiving the anomaly pre-detection signal uploaded by the collection end and reserving the corresponding computing power quota for the corresponding Kanban partition, in order to establish a collaborative contract link of bidirectional constraints between sampling coding and Kanban computing power include: Based on the baseline quota in the computing power quota profile of each partition, the upper limit of the encoding level and the data transmission rate of the corresponding acquisition channel are deduced in reverse, a hierarchical encoding contract is generated and sent to the edge acquisition unit. The receiver receives an anomaly pre-detection signal identified by the trend summary layer, allocates a preset proportion of computing power from the adjustable margin of the corresponding partition as reserved computing power, and updates the temporary coding permissions in the hierarchical coding contract. A dedicated transmission channel for binding coding contracts and computing power quotas is established between the acquisition terminal and the corresponding view partition of the dashboard, forming the collaborative contract link.

4. The collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard according to claim 3, characterized in that, The steps of constructing independent alarm feature mirrors at the data acquisition end and the dashboard end respectively, and performing difference verification on the dual-end alarm feature mirrors through the collaborative contract link to generate a valid alarm set with consistent verification include: The acquisition end extracts time-series fluctuation features based on the original full-volume data and constructs an alarm feature mirror on the acquisition side; the dashboard end extracts view aggregation features based on the received feature aggregation layer data and constructs an alarm feature mirror on the dashboard side. When an alarm candidate is triggered at either end, the alarm feature image of this end is synchronized to the other end through the collaborative contract link, and the feature dimension difference and temporal offset of the feature images of the two ends are calculated. When both the difference and the offset are within the preset allowable range, the alarm candidate is marked as a valid alarm and included in the valid alarm set.

5. The collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard according to claim 4, characterized in that, The step of dynamically redrawing the computing power partition boundaries of each view partition of the dashboard based on the effective alarm set includes: Collect the view partitions and alarm levels corresponding to each alarm in the valid alarm set, and calculate the computing power increment required for each partition; According to a preset order, computing power is borrowed from the adjustable margin of non-alarm-related partitions and added to the alarm-related partitions. Based on the results of the computing power replenishment, the computing power sharding boundaries of each partition are redefined, the partition computing power quota profile is updated, and the updated quota information is synchronized to the collaborative contract link.

6. The collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard according to claim 5, characterized in that, The step of triggering source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set to generate source tracing encrypted data stream includes: Using the occurrence time of a valid alarm and the associated device topology as anchor points, determine the time window and device link range for source tracing and rollback; An encrypted sampling command is sent to the corresponding acquisition unit to raise the sampling granularity of the device link within the time window from the current level to the original full volume level, and to collect the missing fine-grained data of the rollback period. The fine-grained data collected is time-aligned and linked to generate the source-tracing encrypted data stream.

7. The collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard according to claim 6, characterized in that, The step of mapping the source-tracing encrypted data stream to a source-tracing-specific view partition of the dashboard to form a source-tracing link that links sampling granularity with view rendering includes: Create a dedicated view partition for tracing the source in the dashboard and allocate an independent computing power quota to this partition; The source-tracing encrypted data stream is mapped to a source-specific view partition according to the device topology and time sequence, generating a time-series curve and topology link diagram with sampling granularity identifier; Establish a linkage between view zooming operations and sampling granularity. When the view zoom ratio changes, dynamically adjust the sampling encoding level for the corresponding time period through the collaborative contract link.

8. A collaborative processing system for industrial equipment data acquisition and intelligent monitoring dashboards, characterized in that, The system includes: The quantization module is used to perform edge-level hierarchical encoding processing on the operating data of industrial equipment, generate a sampling encoding stream containing multi-level sampling granularity, and perform rendering computing power pool sharding quantization processing on each view partition of the intelligent monitoring dashboard to generate a partition computing power quota profile. The construction module is used to reverse generate the hierarchical coding contract of the collection terminal based on the partition computing power quota profile, synchronously receive the abnormal pre-detection signal uploaded by the collection terminal and reserve the computing power quota of the corresponding Kanban partition for it, so as to establish a collaborative contract link of bidirectional constraints between sampling coding and Kanban computing power. The verification module is used to build independent alarm feature images at the acquisition end and the dashboard end respectively, perform difference verification on the alarm feature images of the two ends through the collaborative contract link, generate a valid alarm set with consistent verification, and dynamically redraw the computing power partition boundaries of each view partition of the dashboard based on the valid alarm set. The linkage module is used to trigger the source tracing rollback encrypted sampling at the acquisition end based on the valid alarm set, generate the source tracing encrypted data stream, and map the source tracing encrypted data stream to the source tracing dedicated view partition of the dashboard to form a source tracing link that links sampling granularity and view rendering.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the collaborative processing method for industrial equipment data acquisition and intelligent monitoring dashboard as described in any one of claims 1 to 7.

Images