New energy industry supply chain monitoring and early warning method and system based on behavior pattern

Abstract

The invention provides a new energy industry supply chain monitoring and early warning method and system based on a behavior pattern, and relates to the technical field of new energy, and the method comprises the steps: configuring K edge behavior recognition units in a new energy industry chain; when the first edge behavior recognition unit detects a first behavior deviation feature, W downstream cascade recognition units are positioned in the supply chain topology; after an early warning risk instruction sent by the first edge behavior recognition unit is responded, W associated behavior deviation characteristics are analyzed and returned; and fusing the first behavior deviation feature and the W associated behavior deviation features, and outputting a visual alarm map. According to the method and the device, the technical problem that the stability of the new energy industry supply chain is affected due to incomplete identification and early warning of potential risks caused by lack of a global view angle in supply chain monitoring in the prior art is solved, and the stability of the new energy industry supply chain is improved by identifying the behavior deviation characteristics of each node in the supply chain and combining downstream cascade identification and cloud fusion. And the stability of a new energy industry supply chain is improved.

Description

Technical Field

[0001] The present application relates to the field of new energy technology, and in particular to a new energy industry supply chain monitoring and early warning method and system based on behavioral patterns. Background Art

[0002] In the field of new energy industry supply chain management, supply chain monitoring methods primarily rely on single-node behavior monitoring or local optimization, typically monitoring the performance of a specific link. However, supply chain operations are complex and dynamic, and the performance of a single link cannot represent the health of the entire supply chain. This makes it difficult to effectively address the interactions between various links, leading to insufficient awareness of the overall risk of the entire supply chain. Furthermore, communication and data flow between different supply chain nodes (such as raw material suppliers, manufacturers, and transporters) are often not smooth. Each node often operates independently and lacks an effective information sharing mechanism. In particular, when risks arise in a single link, they can have a chain reaction throughout the entire supply chain. However, due to the lack of a cross-node risk identification mechanism, the opportunity to take preventative measures can be missed, resulting in delayed emergency responses and ultimately impacting the stability of the entire supply chain.

[0003] In summary, there are technical problems in the existing technology due to the lack of a global perspective in supply chain monitoring, which leads to incomplete identification and early warning of potential risks, thus affecting the stability of the new energy industry supply chain. Summary of the Invention

[0004] The purpose of this application is to provide a new energy industry supply chain monitoring and early warning method and system based on behavioral patterns, in order to solve the technical problem in the existing technology that the supply chain monitoring lacks a global perspective, resulting in incomplete identification and early warning of potential risks, thereby affecting the stability of the new energy industry supply chain.

[0005] In view of the above problems, this application provides a new energy industry supply chain monitoring and early warning method and system based on behavioral patterns.

[0006] In the first aspect, the present application provides a new energy industry supply chain monitoring and early warning method based on behavioral patterns, and the new energy industry supply chain monitoring and early warning method based on behavioral patterns is implemented through a new energy industry supply chain monitoring and early warning system based on behavioral patterns, wherein the new energy industry supply chain monitoring and early warning method based on behavioral patterns includes: configuring K edge behavior identification units at K supply chain nodes in the new energy industry chain; when the first edge behavior identification unit detects that there is a first behavior deviation feature at the first supply chain node, positioning W downstream cascade identification units in the supply chain topology with the first supply chain node as the warning risk starting point; after the W downstream cascade identification units respond to the warning risk instruction sent by the first edge behavior identification unit, they analyze and return W related behavior deviation features to the supply chain monitoring cloud; the supply chain monitoring cloud receives and integrates the first behavior deviation feature and W related behavior deviation features, and outputs a visual alarm map.

[0007] Optionally, the periodic behavior patterns of the K supply chain nodes are tracked to establish K groups of periodic behavior baselines; after configuring the K edge behavior identification units for the K supply chain nodes, the K groups of periodic behavior baselines are used to localize anomaly detection rules; based on the multivariate transmission records of the K supply chain nodes, a supply chain topology is constructed; and based on the node out-degree direction of the supply chain topology, a cascade communication link of the K edge behavior identification units is established.

[0008] Optionally, the first historical behavior data of the first supply chain node is called locally; the first historical behavior data is divided into multiple first behavior data segments based on a standardized segmentation period; after performing new energy sensitive target tracking on the multiple first behavior data segments, triple exponential smoothing processing is performed on the tracking results to generate multiple segment behavior baselines; after time series splicing and restoring the multiple segment behavior baselines, periodic change feature fitting is performed to obtain W periodic behavior baselines corresponding to W connection periods as the first group of periodic behavior baselines.

[0009] Optionally, based on the first periodic characteristics of the first supply chain node, a first real-time behavior baseline is extracted from the first group of periodic behavior baselines; the first edge behavior identification unit performs real-time behavior detection of the first supply chain node based on the first real-time behavior baseline; if the detection identifies a first behavior deviation feature, a first risk conduction coefficient of the first behavior deviation feature is calculated based on the first real-time behavior baseline; the first conduction diffusion scale is matched according to the first risk conduction coefficient; with the first supply chain node as the starting point for warning risk, the W downstream cascade identification units are positioned in the supply chain topology according to the first conduction diffusion scale.

[0010] Optionally, a first baseline median, a first fluctuation upper limit, and a first fluctuation lower limit are extracted from the first real-time behavior baseline; the first behavior deviation feature is used to traverse the first baseline median, the first fluctuation upper limit, and the first fluctuation lower limit, and after calculating the first deviation rate sequence, the first deviation rate sequence is averaged and a first absolute deviation rate is output; the first deviation weight and the first critical weight of the first supply chain node are called from the industry transmission factor library; the first absolute deviation rate is processed using the first deviation weight and the first critical weight, and the first risk transmission coefficient is output.

[0011] Optionally, the first edge behavior identification unit encapsulates the first risk transmission coefficient and the first behavior deviation feature into a structured warning package; the first edge behavior identification unit sends the structured warning package as a warning risk instruction in parallel to the W downstream cascade identification units through a cascade communication link; after the W downstream cascade identification units respond to the warning risk instruction, they analyze and return the W associated behavior deviation features to the supply chain monitoring cloud.

[0012] Optionally, after time-series alignment of the first behavior deviation feature and W associated behavior deviation features, the first behavior deviation feature and W associated behavior deviation features are mapped to the supply chain topology to obtain a local time-series risk topology; dynamic deduction of the risk transmission path is performed on the local time-series risk topology to generate a risk transmission link, wherein the risk transmission link carries a risk transmission timestamp; after risk level matching of the local time-series risk topology based on the risk deviation feature, three-dimensional thermal rendering is performed based on the node risk level time-series feature to output a risk rendering heat map; the risk transmission link is dynamically superimposed on the risk rendering heat map to output the visual alarm map.

[0013] In the second aspect, the present application also provides a new energy industry supply chain monitoring and early warning system based on behavioral patterns, which is used to execute the new energy industry supply chain monitoring and early warning method based on behavioral patterns as described in the first aspect, wherein the new energy industry supply chain monitoring and early warning system based on behavioral patterns includes: an edge deployment module, which is used to configure K edge behavior identification units at K supply chain nodes in the new energy industry chain; a first deviation identification module, which is used to locate W downstream cascade identification units in the supply chain topology with the first supply chain node as the warning risk starting point when the first edge behavior identification unit detects that there is a first behavior deviation feature in the first supply chain node; a downstream cascade module, which is used for the W downstream cascade identification units to analyze and return W related behavior deviation features to the supply chain monitoring cloud after responding to the warning risk instruction sent by the first edge behavior identification unit; a deviation feature fusion module, which is used for the supply chain monitoring cloud to receive and fuse the first behavior deviation feature and W related behavior deviation features, and output a visual alarm map.

[0014] One or more technical solutions provided in this application have at least the following beneficial effects: By configuring K edge behavior recognition units at K supply chain nodes in the new energy industry chain; when the first edge behavior recognition unit detects the presence of a first behavior deviation feature at the first supply chain node, W downstream cascade recognition units are positioned in the supply chain topology with the first supply chain node as the starting point for the early warning risk; after the W downstream cascade recognition units respond to the early warning risk instruction sent by the first edge behavior recognition unit, they analyze and return W related behavior deviation features to the supply chain monitoring cloud; the supply chain monitoring cloud receives and fuses the first behavior deviation feature and the W related behavior deviation features, and outputs a visual alarm map. In other words, through distributed edge computing units, the behavior deviation features of each node in the supply chain are monitored and identified in real time, and the early warning information is cascaded to the downstream nodes, the behavior deviation features associated with them are analyzed, and the information is integrated in the cloud to output a visual alarm map, providing a global perspective of risk warnings, quickly responding to abnormal situations in the supply chain, and thus improving the stability and efficiency of the new energy industry supply chain.

[0015] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0017] Figure 1 This is a flow chart of the new energy industry supply chain monitoring and early warning method based on behavioral patterns in this application.

[0018] Figure 2 This is a structural diagram of the new energy industry supply chain monitoring and early warning system based on behavioral patterns in this application.

[0019] Explanation of the reference numerals: edge deployment module 11 , first deviation identification module 12 , downstream cascade module 13 , deviation feature fusion module 14 . DETAILED DESCRIPTION

[0020] This application solves the technical problem in the existing technology of incomplete identification and early warning of potential risks due to the lack of a global perspective in supply chain monitoring, thereby affecting the stability of the new energy industry supply chain by providing a monitoring and early warning method and system for the new energy industry supply chain based on behavioral patterns. Through distributed edge computing units, the behavioral deviation characteristics of each node in the supply chain are monitored and identified in real time, and the early warning information is cascaded to downstream nodes. The associated behavioral deviation characteristics are analyzed, integrated in the cloud, and a visual alarm map is output. This provides a global perspective of risk warning and quickly responds to abnormal situations in the supply chain, thereby improving the stability and efficiency of the new energy industry supply chain.

[0021] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0022] For example 1, please refer to the attached Figure 1 The present application provides a new energy industry supply chain monitoring and early warning method based on behavioral patterns, wherein the new energy industry supply chain monitoring and early warning method based on behavioral patterns is executed by a new energy industry supply chain monitoring and early warning system based on behavioral patterns, and the new energy industry supply chain monitoring and early warning method based on behavioral patterns specifically includes the following steps: S100: Configure K edge behavior recognition units at K supply chain nodes in the new energy industry chain.

[0023] Furthermore, the present application S100 includes: The periodic behavior patterns of the K supply chain nodes are tracked to establish K groups of periodic behavior baselines. After configuring the K edge behavior identification units for the K supply chain nodes, the K groups of periodic behavior baselines are used to localize anomaly detection rules. A supply chain topology is constructed based on the multivariate transmission records of the K supply chain nodes. A cascade communication link of the K edge behavior identification units is established according to the node out-degree direction of the supply chain topology.

[0024] Furthermore, the present application further comprises the following steps: The first historical behavior data of the first supply chain node is locally called; the first historical behavior data is divided into multiple first behavior data segments based on a standardized segmentation period; after performing new energy sensitive target tracking on the multiple first behavior data segments, triple exponential smoothing processing is performed on the tracking results to generate multiple segment behavior baselines; after time series splicing and restoring the multiple segment behavior baselines, periodic change feature fitting is performed to obtain W periodic behavior baselines corresponding to W connection periods as the first group of periodic behavior baselines.

[0025] Specifically, in supply chain management, a node typically refers to each link or entity in the new energy industry supply chain, specifically, multiple key nodes selected within the new energy industry supply chain. A node is randomly selected from K supply chain nodes as the first supply chain node, representing any link in the supply chain. Historical behavioral data for the first supply chain node is obtained, typically including various operational data from the first supply chain node over a period of time, such as production data, delivery data, inventory data, and transportation data, for analysis of its performance.

[0026] Segment the first historical behavior data according to a standardized segmentation period to obtain multiple first behavior data segments. A standardized segmentation period divides data into fixed time intervals, ensuring that the lengths of different data segments are consistent. For example, if the data is for a month, you can segment it by week or day, ensuring that the data length of each time period is consistent.

[0027] New energy sensitive target tracking is performed on multiple first-behavior data segments. Data analysis identifies whether supply chain nodes are meeting key performance indicators. New energy sensitive target tracking involves monitoring and tracking key objectives related to the new energy industry (such as energy efficiency, output, and quality) to determine whether the node's behavior meets expected standards or goals. The tracking results are then triple-exponentially smoothed to remove noise and highlight key trends and cyclical changes.

[0028] Triple exponential smoothing is a time series smoothing technique used to remove noise from data and identify trends and seasonality. It uses three different weighted averages to smooth the data, allowing for more accurate predictions of future trends. Triple exponential smoothing involves smoothing the data at three different levels: a smoothing constant (used to smooth the trend of the data), a seasonal index (used to smooth cyclical fluctuations), and a trend index (used to smooth long-term trends). Horizontal smoothing, by calculating a weighted average of the current and previous values, is primarily used to remove random fluctuations. Trend smoothing is used to smooth long-term trends, helping to identify long-term trends in production or other behaviors. Seasonal smoothing is used to capture cyclical changes, such as seasonal fluctuations in demand, and then adjust the data to reduce the impact of seasonal fluctuations.

[0029] Through triple exponential smoothing, we generate multiple segmented behavioral baselines. Each segmented baseline represents the normal behavior pattern of a supply chain node within that period, such as normal production cycles and inventory fluctuations. It demonstrates the typical performance of a supply chain node within a specific time period. Each segmented baseline can be used to determine whether the node deviates from its normal behavior pattern. If the actual data for a given time period deviates significantly from the baseline, this indicates abnormal behavior at that node, potentially indicating potential risk.

[0030] Multiple segmented behavioral baselines are reassembled to restore a complete historical behavioral time series, essentially concatenating all data segments into a single long time series. Based on the assembled complete baselines, cyclical variation characteristics are fitted to identify cyclical behavioral patterns of supply chain nodes, including seasonal variations, production peaks, and demand fluctuations. Based on the fitting results (i.e., cyclical variation behavior patterns), W cyclical behavior baselines corresponding to W connected periods are obtained, representing the behavioral patterns of supply chain nodes in different time periods and serving as the first set of cyclical behavior baselines. For example, the historical behavioral data of nodes in the new energy industry chain is segmented into 12 data segments by month, and each segment is subjected to triple exponential smoothing to obtain 12 segmented behavior baselines. These 12 segmented behavior baselines are then reassembled in chronological order to form a complete behavioral pattern. It is found that production efficiency is higher in the summer (June, July, and August) and lower in the winter (December, January, and February). Therefore, cyclical variation characteristics are fitted to establish four cyclical behavior baselines corresponding to four connected periods, representing the production efficiency patterns of spring, summer, autumn, and winter, respectively.

[0031] Repeat the above process for K supply chain nodes to establish K sets of cyclical behavioral baselines. By generating these cyclical behavioral baselines, we can accurately identify the normal behavior patterns of each supply chain node and uncover potential anomalies. Triple exponential smoothing effectively removes noise and short-term fluctuations from the data, highlighting long-term trends and providing a more stable behavioral baseline. By tracking multiple cyclical behavioral patterns, we can clearly identify the changing trends of supply chain nodes over different time periods, thereby improving the efficiency and stability of supply chain management.

[0032] Configure K edge behavior recognition units for K supply chain nodes. This means deploying an edge computing unit at each supply chain node, enabling each node to independently collect, process, and analyze data. An edge behavior recognition unit is an edge computing device or system unit typically deployed at each supply chain node. It monitors, analyzes, and identifies behavioral data anomalies in real time, possessing a certain level of local computing power for rapid data response and processing.

[0033] Each edge behavior recognition unit generates and applies corresponding anomaly detection rules based on the corresponding periodic behavior baseline. Specifically, the edge computing unit will analyze the real-time data of each node to detect whether there is any behavior that deviates significantly from the predetermined baseline and determine whether an anomaly has occurred. Localization of anomaly detection rules refers to the use of local data analysis and behavior pattern baselines in the edge computing unit to formulate and execute anomaly detection rules. This means that anomaly detection for each node will be processed locally based on its historical behavior and periodic patterns, without relying on cloud computing.

[0034] Obtaining multi-source transmission records for K supply chain nodes—that is, recording and transmitting data links between multiple supply chain nodes—helps us understand the behavioral connections and interactive impacts between individual nodes. In other words, recording the data transmission relationships between supply chain nodes allows us to understand the data flows and dependencies between different nodes. Based on these multi-source transmission records, we construct a topological diagram of the entire supply chain. Supply chain topology helps identify the connections between nodes, thereby clarifying the paths of data and material flow. Supply chain topology refers to the structural relationships between nodes in the supply chain, including how they connect, interact, and how data and resources flow between them.

[0035] Cascade communication links are established between nodes based on the node outdegree direction of the supply chain topology. For example, if a node's outdegree points to multiple downstream nodes, information is transmitted sequentially, ensuring smooth information flow within the supply chain. Cascade communication links ensure that data and instructions can be passed between different nodes layer by layer, effectively coordinating all links in the supply chain. A node's outdegree direction is the number and direction of a node's outgoing edges, indicating the direction in which information or resources from that node flow to other nodes. By establishing communication links between nodes, information can be transmitted layer by layer within the supply chain. Cascade communication links typically represent dependencies between nodes and are used to synchronize data between nodes.

[0036] Edge computing units perform anomaly detection locally at each supply chain node, enabling real-time monitoring and rapid response to supply chain anomalies, reducing latency associated with cloud-based reliance. By building a supply chain topology and designing cascaded communication links, we ensure that information and instructions are delivered quickly and accurately across all links in the supply chain, optimizing supply chain management and scheduling.

[0037] S200: When a first edge behavior recognition unit detects a first behavior deviation feature in a first supply chain node, W downstream cascade recognition units are located in the supply chain topology with the first supply chain node as a warning risk starting point.

[0038] Furthermore, the present application S200 includes: According to the first periodic characteristics of the first supply chain node, a first real-time behavior baseline is extracted from the first group of periodic behavior baselines; the first edge behavior identification unit performs real-time behavior detection on the first supply chain node based on the first real-time behavior baseline; if a first behavior deviation feature is detected and identified, a first risk conduction coefficient of the first behavior deviation feature is calculated based on the first real-time behavior baseline; a first conduction diffusion scale is matched according to the first risk conduction coefficient; with the first supply chain node as the starting point for early warning risk, the W downstream cascade identification units are positioned in the supply chain topology according to the first conduction diffusion scale.

[0039] Furthermore, the present application further comprises the following steps: Extract the first baseline median, the first fluctuation upper limit and the first fluctuation lower limit from the first real-time behavior baseline; use the first behavior deviation feature to traverse the first baseline median, the first fluctuation upper limit and the first fluctuation lower limit, calculate the first deviation rate sequence, perform mean calculation on the first deviation rate sequence, and output the first absolute deviation rate; call the first deviation weight and the first critical weight of the first supply chain node from the industry transmission factor library; use the first deviation weight and the first critical weight to process the first absolute deviation rate, and output the first risk transmission coefficient.

[0040] Specifically, a first cycle feature of the first supply chain node is obtained, i.e., a periodic feature related to the first supply chain node, such as a production cycle, inventory cycle, or delivery cycle. Based on the first cycle feature, a first real-time behavior baseline is randomly extracted from the first set of periodic behavior baselines. The first real-time behavior baseline reflects the behavior of the first supply chain node in the first cycle. The first edge behavior recognition unit of the first supply chain node performs real-time behavior detection on the first real-time behavior baseline to detect whether the behavior of the first supply chain node conforms to a normal pattern and identify any deviations.

[0041] If, after real-time behavior monitoring, the detection identifies the presence of a first behavioral deviation characteristic, that is, there is a difference between the real-time behavior data and its standard behavior pattern (the first real-time behavior baseline), the median value, upper and lower fluctuation limits are extracted from the first real-time behavior baseline, representing the standard range of the first supply chain node behavior. Any data outside this range will be considered abnormal. The baseline median is the middle value in a baseline data set, indicating the typical value in the node behavior pattern, and is usually used to measure the central tendency of the data. The upper and lower fluctuation limits refer to the upper and lower limits of the fluctuation range, respectively, and are used to determine the normal fluctuation range of the data. Exceeding this range indicates that an abnormality may have occurred.

[0042] The first behavioral deviation feature is used to traverse the first baseline median, the first upper fluctuation limit, and the first lower fluctuation limit to calculate the first deviation rate sequence. The first behavioral deviation feature is compared with the baseline median and the fluctuation range (upper and lower limits) to determine whether each data point falls within the normal range. If the first behavioral deviation feature is greater than the upper fluctuation limit, the behavior is deviating positively. If the first behavioral deviation feature is less than the lower fluctuation limit, the behavior is deviating negatively. If the first behavioral deviation feature is within the fluctuation range, the behavior is normal. For example, if the first baseline median is 100, the first upper fluctuation limit is 110, and the first lower fluctuation limit is 90, any behavior outside this range is considered abnormal. If the first behavioral deviation feature is 105, the first median deviation rate is (105-100) / 100 = 5%, the first upper limit deviation rate is (110-105) / 110 = 4.5%, and the first lower limit deviation rate is (105-90) / 90 = 16.6%. The first deviation rate sequence is a sequence of comparisons between the deviation characteristics of the first behavior and the median value of the first baseline, the first upper fluctuation limit, and the first lower fluctuation limit. The first absolute deviation rate is calculated by averaging the first deviation rate sequence, representing the overall degree of deviation in the behavior of the first supply chain node. The first absolute deviation rate is calculated by dividing the sum of all deviation rates in the first deviation rate sequence by the number of data points. This represents the overall degree of deviation of the node from the baseline during that period and indicates the overall change in the node's behavior during the monitoring period.

[0043] Based on the type of the first supply chain node, the first deviation weight and the first criticality weight are called from the industry transmission factor library to measure the potential impact of the node's deviation behavior on the entire supply chain. The industry transmission factor library is a database used to store information such as the behavioral deviation, criticality, and weight of each node in the supply chain. It provides each supply chain node with relevant risk transmission coefficients, weights, and other data for calculating and assessing potential risks. The deviation weight indicates the degree of impact of a node's behavioral deviation on the entire supply chain system. Nodes with higher deviation weights have a greater impact on the supply chain due to their deviation behavior. The criticality weight is the importance of a node in the supply chain, usually measured by the node's contribution to the overall stability and efficiency of the supply chain.

[0044] The first absolute deviation rate is processed based on the first deviation weight and the first criticality weight to produce a risk transmission coefficient, reflecting the potential impact of a node's abnormal behavior on the stability and operation of the entire supply chain. The higher the risk transmission coefficient, the more likely the deviant behavior at that node is to trigger supply chain risks. The first risk transmission coefficient is an indicator of the risk transmission capacity of the first supply chain node, output by comprehensively considering the first absolute deviation rate, the first deviation weight, and the first criticality weight. The first risk transmission coefficient is used to measure the potential risk impact of the deviant behavior of the first supply chain node on the entire supply chain.

[0045] Based on the first risk transmission coefficient, an appropriate first transmission diffusion scale is assigned to reflect the scope of impact of a node's deviation in the supply chain. For example, a deviation at a key manufacturing node could have a wide-ranging impact on multiple downstream nodes, so its transmission diffusion scale is large. The first transmission diffusion scale indicates the scope and impact of risk in the supply chain. The larger the transmission diffusion scale, the wider the impact of the node's risk on downstream nodes and the entire supply chain.

[0046] Taking the first supply chain node as the starting point for early warning risk, the system locates W downstream cascade identification units in the supply chain topology based on the first transmission and diffusion scale to determine which downstream nodes in the supply chain topology may be affected. By analyzing the topology, W potentially affected downstream nodes are located and the corresponding risk identification units are activated. Edge computing units monitor node behavior in real time and, combined with baseline analysis, quickly detect deviations and calculate the corresponding risk transmission coefficients, issuing early warnings. Based on the risk transmission coefficients and transmission and diffusion scales, the system accurately assesses the spread of risks, ensuring that relevant nodes in the supply chain can respond promptly to potential risks. By locating affected downstream nodes in the supply chain topology, responses can be effectively coordinated across various links, mitigating the risk's impact on the entire supply chain.

[0047] S300: After responding to the early warning risk instruction sent by the first edge behavior recognition unit, the W downstream cascade recognition units analyze and return W associated behavior deviation features to the supply chain monitoring cloud.

[0048] Furthermore, the present application S300 includes: The first edge behavior identification unit encapsulates the first risk transmission coefficient and the first behavior deviation feature into a structured warning package; the first edge behavior identification unit sends the structured warning package as a warning risk instruction in parallel to the W downstream cascade identification units through a cascade communication link; after the W downstream cascade identification units respond to the warning risk instruction, they analyze and return the W associated behavior deviation features to the supply chain monitoring cloud.

[0049] Specifically, the first edge behavior identification unit encapsulates the first behavior deviation characteristics and the calculated first risk transfer coefficient into a structured warning package. This package is stored in a standardized format (such as JSON or XML) for easy transmission and parsing. A structured warning package is a data packet format used to encapsulate risk warning information. This package encapsulates key information, such as the first risk transfer coefficient and the first behavior deviation characteristics, to ensure information integrity and effective transmission.

[0050] Through a cascade communication link, the first edge behavior recognition unit transmits the structured warning package to downstream nodes, namely, W downstream cascade recognition units. Parallel transmission means that multiple downstream recognition units can simultaneously receive and independently process the warning package, improving response speed and ensuring timely information delivery to each downstream node. A cascade communication link is a chain-like connection established between supply chain nodes through network communication. Risk information from one node can be transmitted to multiple downstream nodes via the cascade communication link, ensuring information dissemination.

[0051] After receiving the early warning risk instruction, each of the W downstream cascade identification units performs anomaly detection based on its own real-time behavioral data compared to the real-time behavioral baseline, analyzing whether there are deviation characteristics similar to those of the upstream node. If an anomaly is found, the analysis results (such as the behavioral deviation characteristics of the node itself) are transmitted back to the supply chain monitoring cloud. Associated behavioral deviation characteristics refer to abnormal behaviors that are related to the behavioral deviation characteristics of the upstream node and are monitored by the downstream node after receiving the early warning. The supply chain monitoring cloud is a cloud platform that centrally manages and analyzes supply chain data. It receives early warning information and behavioral data from each node and conducts summary analysis.

[0052] Through edge computing units and cascaded communication links, early warning information in the supply chain is rapidly transmitted to all relevant downstream nodes, reducing information transmission delays. Downstream nodes can perform behavioral analysis upon receiving early warning information and promptly transmit deviation characteristic data, further enhancing the supply chain's real-time monitoring capabilities. By aggregating data transmitted back from downstream nodes to the cloud, comprehensive analysis of the risk status of the entire supply chain is possible, enabling timely identification and response to abnormal behavior at each node, effectively mitigating the impact of potential risks on the entire supply chain and enhancing its resilience and stability.

[0053] S400: The supply chain monitoring cloud receives and fuses the first behavior deviation feature and W associated behavior deviation features, and outputs a visual alarm map.

[0054] Furthermore, the present application S400 includes: After time-series alignment of the first behavior deviation feature and W associated behavior deviation features, the first behavior deviation feature and the W associated behavior deviation features are mapped to the supply chain topology to obtain a local time-series risk topology; dynamic deduction of the risk transmission path is performed on the local time-series risk topology to generate a risk transmission link, wherein the risk transmission link carries a risk transmission timestamp; after risk level matching of the local time-series risk topology based on the risk deviation feature, three-dimensional thermal rendering is performed based on the node risk level time-series feature to output a risk rendering heat map; the risk transmission link is dynamically superimposed on the risk rendering heat map to output the visual alarm map.

[0055] Specifically, the behavioral deviation features of all relevant nodes within the K supply chain nodes (i.e., the first supply chain node and its W associated behavioral deviation features) are aligned in chronological order to ensure that behavioral deviations at the same point in time across different nodes can be correctly compared and analyzed. Temporal alignment involves aligning the behavioral deviation features of multiple nodes in chronological order, allowing deviation features from different time points to be integrated and analyzed uniformly.

[0056] The first behavior deviation signature and W associated behavior deviation signatures after time series alignment are mapped to the supply chain topology to obtain a local time series risk topology, which displays the risk profile of each node and their relationships within a specific time period. A local time series risk topology is a risk topology generated based on the supply chain topology, node behavior data, and time series alignment. It maps the behavior deviation signatures of multiple supply chain nodes to the supply chain topology in chronological order, forming a localized time series risk map.

[0057] Within the local time-series risk topology, dynamic risk transmission paths are simulated and calculated. By simulating and calculating the process of risk transmission from one node to other nodes, the risk propagation path and potential spread are dynamically deduced based on the risk transmission coefficient and time factor. The risk transmission path from one node (e.g., a factory) to other nodes (e.g., warehouses, distributors, etc.) is simulated and predicted, determining the order and scope of risk transmission. Each transmission link includes a timestamp indicating the time of risk transmission at each node. A risk transmission link is a chain of risk transmission that indicates the path of risk transmission from upstream nodes to downstream nodes, including the time and process of risk transmission. The risk transmission timestamp marks the time nodes in the risk transmission link, reflecting the time required for risk to propagate from one node to another.

[0058] Based on the risk deviation characteristics, the local temporal risk topology is matched to risk levels, and the risk level of each node is assessed. If a node has a large deviation characteristic, the risk level of that node may be higher. The risk level of each node is divided into low risk, medium risk, and high risk based on the severity of its deviation and the potential impact on downstream nodes. Based on the risk level and temporal characteristics of each node, a three-dimensional thermal rendering is performed, using color depth, shape, and size to represent different risk levels, allowing managers to intuitively view the risk status of each node. For example, red indicates high risk, yellow indicates medium risk, and green indicates low risk.

[0059] The risk transmission links derived from the risk transmission path deduction are superimposed on the risk rendering heat map to generate a visual alarm map. The visual alarm map not only displays the risk level of each node, but also shows how risks are transmitted from one node to another, helping managers better identify potential risk transmission chains and take timely countermeasures. Through the dynamic deduction of risk transmission paths, it is clear how risks are transmitted from one node to other nodes, and a comprehensive risk assessment is conducted for each node in the supply chain, considering not only the deviation of a single node but also the interaction effects of each node. Nodes of different risk levels are color-coded to help managers quickly identify potential problems, thereby improving emergency response capabilities and the stability of the new energy industry supply chain.

[0060] In summary, the new energy industry supply chain monitoring and early warning method based on behavioral patterns provided by this application has the following beneficial effects: By configuring K edge behavior recognition units at K supply chain nodes in the new energy industry chain; when the first edge behavior recognition unit detects the presence of a first behavior deviation feature at the first supply chain node, W downstream cascade recognition units are positioned in the supply chain topology with the first supply chain node as the starting point for the early warning risk; after the W downstream cascade recognition units respond to the early warning risk instruction sent by the first edge behavior recognition unit, they analyze and return W related behavior deviation features to the supply chain monitoring cloud; the supply chain monitoring cloud receives and fuses the first behavior deviation feature and the W related behavior deviation features, and outputs a visual alarm map. In other words, through distributed edge computing units, the behavior deviation features of each node in the supply chain are monitored and identified in real time, and the early warning information is cascaded to the downstream nodes, the behavior deviation features associated with them are analyzed, and the information is integrated in the cloud to output a visual alarm map, providing a global perspective of risk warnings, quickly responding to abnormal situations in the supply chain, and thus improving the stability and efficiency of the new energy industry supply chain.

[0061] Example 2, based on the same inventive concept as the new energy industry supply chain monitoring and early warning method based on behavioral patterns in the aforementioned Example 1, this application also provides a new energy industry supply chain monitoring and early warning system based on behavioral patterns, please refer to the attached Figure 2 The new energy industry supply chain monitoring and early warning system based on behavioral patterns includes: The edge deployment module 11 is used to configure K edge behavior recognition units at K supply chain nodes in the new energy industry chain; the first deviation recognition module 12 is used to locate W downstream cascade recognition units in the supply chain topology with the first supply chain node as the starting point for early warning risk when the first edge behavior recognition unit detects the presence of a first behavior deviation feature in the first supply chain node; the downstream cascade module 13 is used to analyze and return W related behavior deviation features to the supply chain monitoring cloud after the W downstream cascade recognition units respond to the early warning risk instruction sent by the first edge behavior recognition unit; the deviation feature fusion module 14 is used to receive and fuse the first behavior deviation feature and W related behavior deviation features on the supply chain monitoring cloud, and output a visual alarm map.

[0062] Furthermore, the edge deployment module 11 in the new energy industry supply chain monitoring and early warning system based on behavior patterns is also used to: The periodic behavior patterns of the K supply chain nodes are tracked to establish K groups of periodic behavior baselines. After configuring the K edge behavior identification units for the K supply chain nodes, the K groups of periodic behavior baselines are used to localize anomaly detection rules. A supply chain topology is constructed based on the multivariate transmission records of the K supply chain nodes. A cascade communication link of the K edge behavior identification units is established according to the node out-degree direction of the supply chain topology.

[0063] Furthermore, the edge deployment module 11 in the new energy industry supply chain monitoring and early warning system based on behavior patterns is also used to: The first historical behavior data of the first supply chain node is locally called; the first historical behavior data is divided into multiple first behavior data segments based on a standardized segmentation period; after performing new energy sensitive target tracking on the multiple first behavior data segments, triple exponential smoothing processing is performed on the tracking results to generate multiple segment behavior baselines; after time series splicing and restoring the multiple segment behavior baselines, periodic change feature fitting is performed to obtain W periodic behavior baselines corresponding to W connection periods as the first group of periodic behavior baselines.

[0064] Furthermore, the first deviation identification module 12 in the new energy industry supply chain monitoring and early warning system based on behavior patterns is further configured to: According to the first periodic characteristics of the first supply chain node, a first real-time behavior baseline is extracted from the first group of periodic behavior baselines; the first edge behavior identification unit performs real-time behavior detection on the first supply chain node based on the first real-time behavior baseline; if a first behavior deviation feature is detected and identified, a first risk conduction coefficient of the first behavior deviation feature is calculated based on the first real-time behavior baseline; a first conduction diffusion scale is matched according to the first risk conduction coefficient; with the first supply chain node as the starting point for early warning risk, the W downstream cascade identification units are positioned in the supply chain topology according to the first conduction diffusion scale.

[0065] Furthermore, the first deviation identification module 12 in the new energy industry supply chain monitoring and early warning system based on behavior patterns is further configured to: Extract the first baseline median, the first fluctuation upper limit and the first fluctuation lower limit from the first real-time behavior baseline; use the first behavior deviation feature to traverse the first baseline median, the first fluctuation upper limit and the first fluctuation lower limit, calculate the first deviation rate sequence, perform mean calculation on the first deviation rate sequence, and output the first absolute deviation rate; call the first deviation weight and the first critical weight of the first supply chain node from the industry transmission factor library; use the first deviation weight and the first critical weight to process the first absolute deviation rate, and output the first risk transmission coefficient.

[0066] Furthermore, the downstream cascade module 13 in the new energy industry supply chain monitoring and early warning system based on behavior patterns is also used to: The first edge behavior identification unit encapsulates the first risk transmission coefficient and the first behavior deviation feature into a structured warning package; the first edge behavior identification unit sends the structured warning package as a warning risk instruction in parallel to the W downstream cascade identification units through a cascade communication link; after the W downstream cascade identification units respond to the warning risk instruction, they analyze and return the W associated behavior deviation features to the supply chain monitoring cloud.

[0067] Furthermore, the deviation feature fusion module 14 in the new energy industry supply chain monitoring and early warning system based on behavior patterns is further used to: After time-series alignment of the first behavior deviation feature and W associated behavior deviation features, the first behavior deviation feature and the W associated behavior deviation features are mapped to the supply chain topology to obtain a local time-series risk topology; dynamic deduction of the risk transmission path is performed on the local time-series risk topology to generate a risk transmission link, wherein the risk transmission link carries a risk transmission timestamp; after risk level matching of the local time-series risk topology based on the risk deviation feature, three-dimensional thermal rendering is performed based on the node risk level time-series feature to output a risk rendering heat map; the risk transmission link is dynamically superimposed on the risk rendering heat map to output the visual alarm map.

[0068] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The new energy industry supply chain monitoring and early warning method based on behavioral patterns and the specific examples in Example 1 are also applicable to the new energy industry supply chain monitoring and early warning system based on behavioral patterns in this embodiment. Through the above detailed description of the new energy industry supply chain monitoring and early warning method based on behavioral patterns, technical personnel in this field can clearly understand the new energy industry supply chain monitoring and early warning system based on behavioral patterns in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0069] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0070] Obviously, for those skilled in the art, several improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the scope of protection of the present application.

Claims

1. A new energy industry supply chain monitoring and early warning method based on behavioral patterns, characterized by: include: Configure K edge behavior recognition units at K supply chain nodes in the new energy industry chain; When the first edge behavior recognition unit detects that the first supply chain node has a first behavior deviation feature, W downstream cascade recognition units are located in the supply chain topology with the first supply chain node as the starting point for early warning risk; After responding to the early warning risk instruction sent by the first edge behavior recognition unit, the W downstream cascade recognition units analyze and return W related behavior deviation features to the supply chain monitoring cloud; The supply chain monitoring cloud receives and fuses the first behavior deviation feature and W associated behavior deviation features, and outputs a visual alarm map.

2. The new energy industry supply chain monitoring and early warning method based on behavioral patterns according to claim 1 is characterized in that: In the new energy industry chain, K supply chain nodes are equipped with K edge behavior recognition units, including: Tracking the periodic behavior patterns of the K supply chain nodes and establishing K groups of periodic behavior baselines; After configuring the K edge behavior recognition units for the K supply chain nodes, localizing anomaly detection rules using the K groups of periodic behavior baselines; Constructing a supply chain topology based on the multivariate transmission records of the K supply chain nodes; A cascade communication link of the K edge behavior identification units is established according to the node out-degree direction of the supply chain topology.

3. The new energy industry supply chain monitoring and early warning method based on behavioral patterns according to claim 1 is characterized in that: When the first edge behavior recognition unit detects that the first supply chain node has a first behavior deviation feature, W downstream cascade recognition units are located in the supply chain topology with the first supply chain node as the starting point for early warning risk, including: extracting a first real-time behavior baseline from a first set of periodic behavior baselines according to a first periodic feature of the first supply chain node; The first edge behavior recognition unit performs real-time behavior detection of the first supply chain node based on the first real-time behavior baseline; If a first behavior deviation feature is detected and identified, calculating a first risk conduction coefficient of the first behavior deviation feature based on the first real-time behavior baseline; matching a first conduction diffusion scale according to the first risk conduction coefficient; Taking the first supply chain node as a starting point for early warning risk, the W downstream cascade identification units are located in the supply chain topology according to the first conduction and diffusion scale.

4. The new energy industry supply chain monitoring and early warning method based on behavioral patterns according to claim 3 is characterized in that: After the W downstream cascade identification units respond to the early warning risk instruction sent by the first edge behavior identification unit, they analyze and return W associated behavior deviation features to the supply chain monitoring cloud, including: The first edge behavior identification unit encapsulates the first risk transmission coefficient and the first behavior deviation feature into a structured early warning package; The first edge behavior recognition unit sends the structured warning package as a warning risk instruction in parallel to the W downstream cascade recognition units through a cascade communication link; After responding to the early warning risk instruction, the W downstream cascade identification units analyze and transmit back the W associated behavior deviation features to the supply chain monitoring cloud.

5. The new energy industry supply chain monitoring and early warning method based on behavioral patterns as claimed in claim 3 is characterized in that: If a first behavioral deviation feature is detected and identified, calculating a first risk conduction coefficient of the first behavioral deviation feature based on the first real-time behavioral baseline includes: extracting a first baseline median value, a first fluctuation upper limit, and a first fluctuation lower limit from the first real-time behavioral baseline; Using the first behavior deviation feature to traverse the first baseline median, the first fluctuation upper limit, and the first fluctuation lower limit, after calculating a first deviation rate sequence, performing mean calculation on the first deviation rate sequence, and outputting a first absolute deviation rate; Retrieving the first deviation weight and the first critical weight of the first supply chain node from the industry transmission factor library; The first absolute deviation rate is processed using the first deviation weight and the first critical weight, and the first risk transmission coefficient is output.

6. The new energy industry supply chain monitoring and early warning method based on behavioral patterns according to claim 1 is characterized in that: The supply chain monitoring cloud receives and fuses the first behavior deviation feature and W associated behavior deviation features, and outputs a visual alarm map, including: After time series alignment of the first behavior deviation feature and the W associated behavior deviation features, the first behavior deviation feature and the W associated behavior deviation features are mapped to the supply chain topology to obtain a local time series risk topology; Performing dynamic deduction of risk transmission paths in the local time-series risk topology to generate risk transmission links, wherein the risk transmission links are provided with risk transmission timestamps; After matching the risk levels of the local temporal risk topology based on the risk deviation characteristics, three-dimensional thermal rendering is performed based on the temporal characteristics of the node risk levels, and a risk rendering heat map is output; The risk transmission link is dynamically superimposed on the risk rendering heat map to output the visual alarm map.

7. The new energy industry supply chain monitoring and early warning method based on behavioral patterns according to claim 1 is characterized in that: Track the periodic behavior patterns of the K supply chain nodes and establish K groups of periodic behavior baselines, including: Locally call the first historical behavior data of the first supply chain node; Segmenting the first historical behavior data into a plurality of first behavior data segments based on a standardized segmentation period; After executing new energy sensitive target tracking on the plurality of first behavior data segments, performing triple exponential smoothing processing on the tracking results to generate a plurality of segment behavior baselines; After the multiple segment behavior baselines are restored by time series splicing, periodic variation feature fitting is performed to obtain W periodic behavior baselines corresponding to W connected periods as the first group of periodic behavior baselines.

8. The new energy industry supply chain monitoring and early warning system based on behavioral patterns is characterized by: Steps for implementing the new energy industry supply chain monitoring and early warning method based on behavioral patterns as described in any one of claims 1 to 7, wherein the new energy industry supply chain monitoring and early warning system based on behavioral patterns comprises: The edge deployment module is used to configure K edge behavior recognition units at K supply chain nodes in the new energy industry chain; A first deviation identification module is configured to, when the first edge behavior identification unit detects a first behavior deviation feature in a first supply chain node, locate W downstream cascade identification units in the supply chain topology with the first supply chain node as a warning risk starting point; A downstream cascade module, configured to analyze and transmit back W associated behavior deviation features to the supply chain monitoring cloud after the W downstream cascade identification units respond to the early warning risk instruction sent by the first edge behavior identification unit; The deviation feature fusion module is used in the supply chain monitoring cloud to receive and fuse the first behavior deviation feature and W related behavior deviation features, and output a visual alarm map.

Images