Industrial energy consumption data compression and storage method and system based on cooperation of multiple edge nodes

Abstract

The invention discloses a multi-edge-node collaborative industrial energy consumption data compression and storage method and system, particularly relates to the technical field of data compression and storage, and realizes hierarchical compression processing by layering industrial energy consumption data acquisition equipment into a main node, a core slave node and a terminal slave node. According to the method, various compression algorithms are used for carrying out lossy or lossless compression on different types of data, and cooperative compression and data redundancy elimination between edge nodes are realized under the coordination of a main node by introducing a distance calculation mechanism of'node hierarchy + auxiliary parameters'. The system constructs a node correlation graph through data abstracts, optimizes a data transmission path based on reinforcement learning, and adopts a cold and hot data classification strategy to improve the storage efficiency. The data compression efficiency and reliability are improved, and the method is suitable for real-time processing and long-term storage of large-scale industrial energy consumption data.

Description

Technical Field

[0001] The present invention relates to the technical field of data compression and storage, and more specifically, to an industrial energy consumption data compression and storage method and system with multi-edge node collaboration. Background Art

[0002] With the development of industrial automation and intelligent manufacturing, the energy consumption data generated in industrial scenarios has shown explosive growth. In large industrial parks or intelligent factories, the equipment is widely distributed, the energy consumption data is complex, and the acquisition frequency is high. The massive data generated poses a huge pressure on the network transmission, edge processing, and storage systems. Existing industrial data compression solutions mostly focus on single-point processing or centralized uploading, lacking hierarchical collaboration and compression strategy adaptive mechanisms, resulting in insufficient data compression accuracy, high system load, and even data redundancy, link congestion, or faulty transmission problems, and unable to meet the requirements of edge intelligent applications for data efficiency and reliability. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide an industrial energy consumption data compression and storage method and system with multi-edge node collaboration to solve the problems raised in the above background art.

[0004] To achieve the above object, the present invention provides the following technical solutions: An industrial energy consumption data compression and storage method with multi-edge node collaboration, comprising the following steps: Classify multiple types of edge computing nodes into master nodes and slave nodes according to their computing and communication capabilities, and the slave nodes include core slave nodes and terminal slave nodes; Dynamically select a model to calculate the node distance based on the node hierarchy and auxiliary parameters; construct a comprehensive computing power model to calculate the remaining computing power of the slave nodes; filter out the slave nodes with remaining computing power lower than the system threshold; generate a data distribution path using reinforcement learning based on the remaining slave nodes; comprehensively obtain a comprehensive score based on the node distance of each data distribution path and the node proportion of each data distribution path, and select the path with the highest comprehensive score as the final distribution path; The master node issues a compression strategy and updates the configuration, and at the same time constructs a system fault tolerance and breakpoint resumption mechanism.

[0005] In a preferred embodiment, the node hierarchy divides the combination relationship between any two nodes into four node hierarchies: between the master node and the core slave node, between the core slave nodes, between the core slave node and the terminal slave node, and other cross-layer node combinations that do not have direct collaborative significance.

[0006] In a preferred embodiment, the auxiliary parameters include data redundancy similarity and link quality.

[0007] In a preferred embodiment, a hierarchy-parameter mapping rule is set, and auxiliary parameters are introduced. When the node hierarchy is between the master node and the core slave node and the link quality is higher than the threshold θ1, the system adopts a distance model weighted by topological hop count and link quality. When the node hierarchy is between core slave nodes and the data redundancy similarity is higher than the threshold θ2, the system adopts a logical distance model that focuses on similarity and combines computing power balance. When the node hierarchy is between the core slave node and the terminal slave node and the link quality is greater than the threshold θ3, the system adopts a combined model that mainly considers geographical distance and link quality. If none of the above conditions are met, it is regarded as a cross-domain invalid cooperation node pair, which is set as non-cooperative or assigned a punitive extremely large distance value. The sum of the distances of each node included in a distribution path represents the node distance of the distribution path.

[0008] In a preferred embodiment, a comprehensive computing power model is constructed to calculate the remaining computing power of the slave node; the remaining computing power of the slave node is calculated by comprehensively considering CPU, GPU, memory, and bandwidth; the core slave node needs to balance local processing and task distribution / forwarding capabilities, so its computing power model emphasizes network and concurrent processing capabilities. The terminal slave node only executes local compression tasks and no longer undertakes forwarding, so its computing power model focuses on evaluating its processing and caching capabilities.

[0009] In a preferred embodiment, slave nodes with remaining computing power lower than the system threshold are filtered out; the system threshold is dynamically adjusted according to the data volume of the distributed data.

[0010] In a preferred embodiment, a comprehensive score is calculated according to the node distance of each data distribution path and the node proportion of each data distribution path through the weighted summation rule.

[0011] In a preferred embodiment, for the node proportion of each data distribution path, first collect the compressed data meta-information uploaded by each slave node; each slave node performs a data digest extraction operation locally, maps its original data into a lightweight feature expression form, and after obtaining the digest information of all slave nodes, constructs a global data correlation graph with the slave nodes as vertices in the graph and the digest similarity as edge weights; the edge weight value is calculated through the digest similarity, and node pairs with low correlation are removed by a threshold strategy, so as to retain the node connections with significant redundancy relationships; then, identify the node pairs with high data correlation in the graph to form a redundant cooperation candidate set; the ratio of the number of slave nodes included in the data distribution path that belong to the redundant cooperation candidate set to the total number of slave nodes in the data distribution path is the node proportion.

[0012] In a preferred embodiment, the system has a fault tolerance and resume interrupted transfer mechanism; during data transmission, a block verification mechanism is adopted to achieve resume interrupted transfer; meanwhile, each edge node is backed up with each other according to the redundant configuration. When a certain node fails, the master node can automatically switch to the backup node for data recovery; all compression, transmission, and decompression operations generate detailed logs, recording time, operation instructions, compression parameters, and abnormal status; after the node goes online again, the system preferentially restores the local cache data and synchronizes the status with the master node through the logs.

[0013] In a preferred embodiment, it includes the following modules: an edge node grading and data acquisition module, a node distance calculation and computing power evaluation module, a reinforcement learning path optimization and scheduling module, and a compression policy distribution and fault tolerance management module. The edge node grading and data acquisition module classifies multiple types of edge computing nodes into master nodes and slave nodes according to the computing and communication capabilities. The slave nodes include core slave nodes and terminal slave nodes. The node distance calculation and computing power evaluation module is used to dynamically calculate the node distance based on the node level and auxiliary parameters; and calculate the remaining computing power of the slave nodes by constructing a comprehensive computing power model. The reinforcement learning path optimization and scheduling module is used to filter out the slave nodes with remaining computing power lower than the system threshold; generate a data distribution path using reinforcement learning according to the remaining slave nodes; comprehensively obtain a comprehensive score based on the node distance of each data distribution path and the node proportion of each data distribution path, and select the path with the highest comprehensive score as the final distribution path. The compression policy distribution and fault tolerance management module is used to distribute the compression policy to the slave nodes through the master node and dynamically update the configuration; support the resume interrupted transfer, node backup, and log recovery mechanisms.

[0014] The technical effects and advantages of the present invention: The present invention realizes an industrial energy consumption data compression mechanism for multi-edge node collaboration by constructing a hierarchical architecture of master nodes, core slave nodes, and terminal slave nodes, effectively alleviating the network and storage pressure brought by data explosion. The system introduces a dynamic node distance calculation method based on "node level + auxiliary parameters", combines multi-dimensional factors such as topological structure, link quality, redundancy similarity, and geographical location, dynamically selects the collaboration path, and uses reinforcement learning to optimize the data distribution decision path, improving the scheduling efficiency and compression performance. At the same time, the master node intelligently generates a compression policy according to the remaining computing power of the nodes and the network status, realizing the adaptive selection and fine control of the compression method, and significantly improving the data processing efficiency and compression accuracy.

[0015] In addition, the present invention constructs a system fault tolerance system including block checksum, breakpoint resumption, mutual backup mechanism and log recovery, enhancing the stability and data security of the system in complex industrial fields. By constructing a redundant cooperation graph through data digest similarity and combining the cold and hot data hierarchical storage strategy, it effectively reduces redundancy, saves bandwidth, and improves the transmission priority and response speed of critical data. The system has good scalability and adaptability, and is particularly suitable for large-scale industrial energy consumption monitoring scenarios with high sampling frequencies, strong node heterogeneity, and complex communication conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] For the convenience of those skilled in the art to understand, the present invention will be further described below in conjunction with the accompanying drawings; Figure 1 It is a schematic flowchart of the industrial energy consumption data compression and storage method for multi-edge node cooperation of the present invention; Figure 2 It is a schematic structural diagram of the industrial energy consumption data compression and storage system for multi-edge node cooperation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

[0018] Embodiment 1: The industrial energy consumption data compression and storage method for multi-edge node cooperation of the present invention, as Figure 1 shown, includes the following steps: In the system initialization stage, first, a comprehensive layout of energy consumption monitoring points is carried out for the factory or industrial site, and edge computing devices are deployed at each key node. According to the differences in data traffic, processing capabilities, and geographical locations, all nodes are divided into the following three categories; Master node: responsible for global scheduling and compression strategy formulation. Core slave node: has computing and relay capabilities and can participate in collaborative compression and data distribution. Terminal slave node: only responsible for data collection and local compression and does not participate in forwarding tasks. Each node is assigned a hierarchical identifier. Among them, the slave nodes include core slave nodes and terminal slave nodes.

[0019] During system initialization, the sampling frequencies of various sensors are configured, corresponding compression algorithms and data processing strategies are specified for different data types, and at the same time, a secure communication mechanism is established to achieve efficient and reliable information interaction between nodes.

[0020] Each slave node collects the energy consumption data collected by local sensors in real time according to a preset sampling period, including parameters such as current, voltage, frequency, power factor, etc., and immediately performs data preprocessing operations locally. This processing process includes using algorithms such as moving window averaging or median filtering to remove instantaneous noise points in the data, and at the same time performing normalization or standardization operations on data from different device sources to ensure unified data formats for subsequent compression and collaborative processing.

[0021] Furthermore, for missing or abnormal values, the system automatically fills them through linear interpolation or historical averaging method, and marks the data integrity status for the compression module to judge whether to participate in compression, so as to achieve rapid cleaning and pre-compression preparation of the original energy consumption data.

[0022] After the preprocessing is completed, the slave node compresses the locally collected data according to time windows. Within each time window, the slave node uses a suitable compression algorithm to compress the data according to the preset data type and compression strategy. For example, for current and voltage data with periodicity or small fluctuations, lightweight lossless compression algorithms such as Swinging Door, Delta coding or Run-Length coding can be used to retain the original data accuracy; while for data with obvious change trends, such as power factor or frequency, lossy compression algorithms based on error tolerance, such as Piecewise Aggregate Approximation (PAA), Symbolic Aggregate approXimation (SAX) or lightweight time series prediction models (such as LSTM prediction residual compression), etc. can be used for compression to effectively reduce the data volume.

[0023] During the compression process, the slave node also adds timestamps, source device numbers, compression parameter information and integrity identifiers to the compressed data, and temporarily stores them in the local cache module. The buffer area is equipped with a high-priority upload queue to ensure that key data (such as abnormal states, energy consumption mutations) are sent to the master node first, guaranteeing the system's fast response ability to important energy consumption information.

[0024] During the compression process, relevant meta-information is generated at the same time, including the start and end timestamps, maximum value, minimum value, average value and the number of original data points of this compression block. These compressed data and meta-information are cached in the high-speed cache area of the edge node locally, and the cache space is automatically managed using the FIFO mechanism. This cache retains the latest period of data before being transmitted to the master node or upper-layer storage, and has a breakpoint resumption and rollback mechanism to handle unexpected interruptions.

[0025] Based on local compression, the edge nodes coordinate through the master node to execute cross-node collaborative compression and redundant data elimination strategies.

[0026] In the actual deployment process, considering the obvious hierarchical differences among edge nodes and the dynamic nature of their network states or data characteristics at different times, this system proposes a dynamic distance calculation mechanism based on "node hierarchy + auxiliary parameters". By combining the hierarchical combination relationship between nodes with auxiliary parameters such as redundancy similarity or network link quality, a suitable node distance calculation model is dynamically selected, so as to more accurately reflect the cost required for collaborative compression or scheduling.

[0027] To enhance the adaptability of distance calculation, an auxiliary parameter is introduced to jointly determine the distance calculation method adopted with the hierarchy. This parameter can be selected according to the application scenario.

[0028] After identifying the hierarchical relationship and auxiliary parameters between nodes, the node distance calculation method to be adopted is dynamically determined according to the specific combination relationship of node pairs and the auxiliary parameters. The system first divides the combination relationship between any two nodes into four node hierarchies: between the master node and the core slave node, between core slave nodes, between the core slave node and the terminal slave node, and other cross-layer node combinations that do not have direct collaborative significance.

[0029] On this basis, a hierarchy-parameter mapping rule is set, and an auxiliary parameter is introduced. The auxiliary parameter includes data redundancy similarity and link quality , which are used to further determine whether the necessary conditions for collaborative compression or path scheduling are met.

[0030] When the combination between nodes is that of the master node and the core slave node, and the link quality is higher than the threshold θ1, the system adopts a distance model weighted by topological hop count and link quality; the formula is as follows: ; where represents the topological hop count, which indicates the number of hops between node i and node j in the network logical topology structure, that is, the number of intermediate nodes required to go from one node to another. It can be obtained in the topological graph through an improved Dijkstra algorithm or Floyd algorithm, reflecting the path length and relay complexity; represents the link quality, which indicates the stability or availability of the communication link between node i and node j; represents the node distance from node i to node j. is a proportional regulation parameter used to flexibly adjust the contribution degree of each index to the total distance according to the actual deployment scenario.

[0031] When they are two core slave nodes, and the data redundancy similarity is higher than the threshold θ2, the system adopts a logical distance model mainly based on similarity and combined with computing power balance; the formula is as follows: ; where Indicates the similarity in data content between node i and node j, which can be calculated through MinHash, SimHash or other digest algorithms. is a proportional regulation parameter used to flexibly adjust the contribution degree of each index to the total distance according to the actual deployment scenario.

[0032] When the node pair is the core slave node and its subordinate terminal slave node, and the link quality is greater than the threshold θ3, the system adopts a combined model mainly based on geographical distance and link quality; the formula is as follows: ; represents the physical space distance between node i and node j; the Euclidean distance can be calculated through the GPS coordinates of the node deployment location, or the distance can be estimated through the wireless signal strength (RSSI). It is mainly used to consider spatial-related factors such as physical communication delay and signal attenuation. is a proportional regulation parameter used to flexibly adjust the contribution degree of each index to the total distance according to the actual deployment scenario.

[0033] If any of the above conditions is not met, it is regarded as a cross-domain invalid collaborative node pair, and is set to be non-collaborative or given a punitive extremely large distance value.

[0034] The sum of the distances of each slave node included in a path represents the node distance of that path.

[0035] The master node first regularly collects the compressed data meta-information uploaded by each slave node (including the core slave node and the terminal slave node). Each slave node performs a data digest extraction operation locally, maps its original data into a lightweight feature expression form, such as a statistical frequency vector, MinHash, SimHash or a feature signature digest, and periodically uploads this digest information to the master node. After obtaining the digest information of all slave nodes, the master node constructs a global data correlation graph with the slave nodes as the vertices in the graph and the digest similarity as the edge weights. The edge weight value is calculated through the digest similarity, and node pairs with low correlation are removed by a threshold strategy, so as to retain the node connections with significant redundancy relationships.

[0036] Then, the master node identifies the node pairs with high data correlation in the graph to form a redundant collaboration candidate set. The ratio of the number of slave nodes included in the data distribution path that belong to the redundant collaboration candidate set to the total number of slave nodes in the data distribution path is the node occupancy ratio.

[0037] By constructing a comprehensive computing power model, the remaining computing power of computing slave nodes such as CPUs, GPUs, memory, and bandwidth is integrated. The core slave nodes need to balance local processing and task distribution / forwarding capabilities. Therefore, their computing power model emphasizes network and concurrent processing capabilities. Terminal slave nodes only execute local compression tasks and no longer undertake forwarding. Their processing and caching capabilities are mainly evaluated. Different weights are assigned to CPUs, GPUs, memory, and bandwidth to achieve different adjustments in the focus of calculating the computing power of slave nodes.

[0038] Filter out slave nodes with remaining computing power lower than the system threshold; the system threshold is dynamically adjusted according to the amount of data that needs to be recompressed. Generate data distribution paths based on the remaining slave nodes using reinforcement learning.

[0039] According to the node distances of each data distribution path and the node ratios of each data distribution path, a comprehensive score is obtained by weighted summation. The specific formula is as follows: Z = ZB * β - D * α +; where Z represents the comprehensive score; D represents the node distance of each data distribution path; ZB represents the node ratio of each data distribution path; α and β are the weight coefficients of the node distance of each data distribution path and the node ratio of each data distribution path.

[0040] Select the path with the highest comprehensive score as the final distribution path.

[0041] After the master node updates the policy, it automatically transmits the new compression parameter configuration to the corresponding slave nodes through the distribution mechanism and continuously tracks the policy adjustment effect, so as to realize the continuous adaptive optimization of the system compression performance in a multi-source heterogeneous environment.

[0042] In addition, the master node pushes the new compression policy to the corresponding slave nodes through the configuration distribution mechanism to ensure that each node continuously optimizes the compression effect according to the actual data changes, so as to realize the adaptive dynamic adjustment of the system compression performance and effectively improve the overall data processing efficiency and quality.

[0043] To ensure that industrial data still has high reliability in the edge node operating environment, this method constructs a complete system fault tolerance and data recovery mechanism. During data transmission, the system uses a block check mechanism to achieve resume from breakpoint and avoid data duplicate upload or data packet loss caused by network interruption. At the same time, each edge node is backed up with each other according to the redundant configuration. When a certain node fails, the master node can automatically switch to the backup node for data recovery. All compression, transmission, and decompression operations generate detailed logs, recording information such as time, operation instructions, compression parameters, and abnormal status, which is convenient for system administrators to track and repair. After the node is back online, the system preferentially restores the local cache data and synchronizes the status with the master node through the logs, ultimately ensuring the consistency and integrity of the compressed data.

[0044] Embodiment 2: The design of the industrial energy consumption data compression and storage system with multi-edge node collaboration in the present invention is based on the method in Embodiment 1, specifically as Figure 2 shown, including the following modules: an edge node classification and data acquisition module, a node distance calculation and computing power evaluation module, a reinforcement learning path optimization and scheduling module, and a compression policy distribution and fault tolerance management module; The edge node classification and data acquisition module classifies multiple types of edge computing nodes into master nodes and slave nodes according to their computing and communication capabilities. The slave nodes include core slave nodes and terminal slave nodes; each slave node collects local energy consumption data and performs preprocessing operations such as noise filtering and normalization; The node distance calculation and computing power evaluation module is used to dynamically calculate the node distance based on the node level and auxiliary parameters; and calculate the remaining computing power of the slave nodes by constructing a comprehensive computing power model; The reinforcement learning path optimization and scheduling module is used to filter out the slave nodes with remaining computing power lower than the system threshold; generate data distribution paths using reinforcement learning according to the remaining slave nodes; comprehensively obtain a comprehensive score based on the node distance of each data distribution path and the node proportion of each data distribution path, and select the path with the highest comprehensive score as the final distribution path; The compression policy distribution and fault tolerance management module is used to distribute compression policies to the slave nodes through the master node and dynamically update the configuration; support mechanisms such as breakpoint resumption, node backup, and log recovery.

[0045] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0046] In several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0047] The unit described as a separation component may or may not be physically separated. The component displayed as a unit may or may not be a physical unit, that is, it may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0048] In addition, each functional unit in various embodiments of the present application may be integrated in a processing unit, may exist separately as individual physical units, or two or more units may be integrated in one unit.

[0049] As described above, only the specific implementation manners of the present application are provided, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered by the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An industrial energy consumption data compression and storage method with multi-edge node collaboration, characterized in that It includes the following steps: Classify multiple types of edge computing nodes into master nodes and slave nodes according to computing and communication capabilities. The slave nodes include core slave nodes and terminal slave nodes; Dynamically select the node distance for model calculation based on the node hierarchy and auxiliary parameters; construct a comprehensive computing power model to calculate the remaining computing power of the slave nodes; filter out the slave nodes with remaining computing power lower than the system threshold; Generate a data distribution path for the remaining slave nodes using reinforcement learning; comprehensively obtain a comprehensive score based on the node distance and node proportion of each data distribution path, and select the path with the highest comprehensive score as the final distribution path; The master node distributes the compression policy and updates the configuration, and at the same time constructs a system fault tolerance and resume mechanism for interrupted transmission.

2. The industrial energy consumption data compression and storage method with multi-edge node collaboration according to claim 1, characterized in that: For the node hierarchy, the combination relationship between any two nodes is divided into four node hierarchies: between the master node and the core slave node, between the core slave nodes, between the core slave node and the terminal slave node, and other cross-layer node combinations that do not have direct collaborative significance.

3. The industrial energy consumption data compression and storage method with multi-edge node collaboration according to claim 1, characterized in that: The auxiliary parameters include data redundancy similarity and link quality.

4. The industrial energy consumption data compression and storage method with multi-edge node collaboration according to claim 3, characterized in that: Set the hierarchy-parameter mapping rule and introduce the auxiliary parameters; When the node hierarchy is between the master node and the core slave node and the link quality is higher than the threshold θ1, the system uses a distance model weighted by topological hop count and link quality; When the node hierarchy is between the core slave nodes and the data redundancy similarity is higher than the threshold θ2, the system uses a logical distance model that focuses on similarity and combines computing power balance; When the node hierarchy is between the core slave node and the terminal slave node and the link quality is greater than the threshold θ3, the system uses a combined model that focuses on geographical location distance and link quality; If none of the above conditions are met, it is regarded as a cross-domain invalid collaborative node pair and is set to be non-collaborative or given a punitive extremely large distance value; The sum of the distances of each node included in a distribution path represents the node distance of the distribution path.

5. The method for compressing and storing industrial energy consumption data with multi-edge node collaboration according to claim 1, characterized in that: Construct a comprehensive computing power model to calculate the remaining computing power of the slave nodes; comprehensively calculate the remaining computing power of the slave nodes based on CPU, GPU, memory, and bandwidth; the core slave node needs to balance local processing and task distribution / forwarding capabilities, so its computing power model focuses on network and concurrent processing capabilities. The terminal slave node only performs local compression tasks and no longer undertakes forwarding, so its computing power model mainly evaluates its processing and caching capabilities.

6. The method for compressing and storing industrial energy consumption data with multi-edge node collaboration according to claim 5, characterized in that: Filter out the slave nodes with remaining computing power lower than the system threshold; the system threshold is dynamically adjusted according to the data volume of the distributed data.

7. The industrial energy consumption data compression and storage method with multi-edge node collaboration according to claim 1, characterized in that: Calculate the comprehensive score according to the node distance and node proportion of each data distribution path through the weighted summation rule.

8. The method for compressing and storing industrial energy consumption data with multi-edge node collaboration according to claim 1, characterized in that: For the node ratio of each data distribution path, first collect the compressed data meta-information uploaded by each slave node; each slave node performs a data digest extraction operation locally, maps its original data into a lightweight feature representation form, and after obtaining the digest information of all slave nodes, constructs a global data correlation graph with the slave nodes as vertices in the graph and the digest similarity as the edge weights; the edge weight values are calculated through the digest similarity, and node pairs with low correlation are removed by a threshold strategy, so as to retain the node connections with significant redundancy relationships; then, identify the node pairs with high data correlation in the graph to form a redundant collaboration candidate set; the ratio of the number of slave nodes included in the data distribution path that belong to the redundant collaboration candidate set to the total number of slave nodes in the data distribution path is the node ratio.

9. The method for compressing and storing industrial energy consumption data with multi-edge node collaboration according to claim 1, characterized in that: The system fault tolerance and resume interrupted transfer mechanism; during data transmission, a block verification mechanism is adopted to achieve resume interrupted transfer; at the same time, each edge node is backed up with each other according to the redundant configuration, and when a certain node fails, the master node can automatically switch to the backup node for data recovery; all compression, transmission and decompression operations generate detailed logs, recording time, operation instructions, compression parameters and abnormal status; after the node is back online, the system preferentially restores the local cache data and synchronizes the status with the master node through the logs.

10. An industrial energy consumption data compression and storage system with multi-edge node collaboration, characterized in that, The compression storage system is based on the method according to any one of claims 1-9, and includes the following modules: an edge node grading and data acquisition module, a node distance calculation and computing power evaluation module, a reinforcement learning path optimization and scheduling module, and a compression policy distribution and fault tolerance management module; The edge node grading and data acquisition module classifies multiple types of edge computing nodes into master nodes and slave nodes according to the computing and communication capabilities, and the slave nodes include core slave nodes and terminal slave nodes; The node distance calculation and computing power evaluation module is used to dynamically calculate the node distance based on the node level and auxiliary parameters; and calculate the remaining computing power of the slave node by constructing a comprehensive computing power model; The reinforcement learning path optimization and scheduling module is used to filter out the slave nodes with the remaining computing power lower than the system threshold; Generate a data distribution path according to the remaining slave nodes by using reinforcement learning; obtain a comprehensive score based on the node distance of each data distribution path and the node ratio of each data distribution path, and select the path with the highest comprehensive score as the final distribution path; The compression policy distribution and fault tolerance management module is used to distribute the compression policy to the slave nodes through the master node and dynamically update the configuration; Support the resume interrupted transfer, node backup and log recovery mechanisms.

Images