A hierarchical operation and maintenance management method and system for wind farms based on multi-level monitoring

Abstract

This application provides a hierarchical operation and maintenance management method and system for wind farms based on multi-level monitoring, relating to the field of operation and maintenance monitoring technology. The method includes: constructing a multi-level monitoring transmission network; performing initial path transmission according to the wind farm operation and maintenance calculation task; triggering a rollback decision mechanism when a failure is detected in the next path node; constructing a set of reachable nodes; evaluating the candidate rollback node set through analysis and screening to obtain a rollback cost index; identifying the first rollback node as a successor node to reconstruct the first jump path, and re-executing the operation and maintenance data transmission of the wind farm operation and maintenance calculation task. This application solves the technical problems in existing technologies where the use of a step-by-step rollback path recovery mechanism when the operation and maintenance request path fails in a multi-level monitoring architecture leads to redundant rollback paths, repeated decisions, multiple path reconstructions, and uncontrollable recovery delays, affecting the efficiency of wind farm operation and maintenance. It achieves rapid self-healing of the global path, improving the efficiency of wind farm operation and maintenance.

Description

Technical Field

[0001] This application relates to the field of operation and maintenance monitoring technology, specifically to a hierarchical operation and maintenance management method and system for wind farms based on multi-level monitoring. Background Technology

[0002] Currently, wind farms generally adopt a multi-level monitoring architecture consisting of an edge layer, a site layer, and a central control layer to achieve remote centralized management and control of operation and maintenance data. In actual operation, when a node in a transmission path fails, causing an operation and maintenance request to fail, path recovery is typically achieved through a step-by-step fallback mechanism. Each time the next hop is detected as unreachable, the process falls back to the previous level node and re-attempts forwarding to the next level. This results in numerous redundant jumps in the fallback path, and each fallback requires repeated execution of fault detection and next-hop selection logic, leading to repeated decision-making and multiple path reconstructions. Because the number of fallback steps is related to the depth of the fault point, the latency of the entire recovery process exhibits uncontrollable fluctuations, making it difficult to meet the wind farm's requirements for real-time transmission of operation and maintenance commands and monitoring data. The lack of a mechanism for determining the optimal fallback node under global reachability constraints results in a large number of redundant detection and fallback operations, wasting network resources. Furthermore, the decision-making process is repetitive and short-sighted, failing to evaluate the optimal recovery point globally, leading to multiple inefficient path reconstruction attempts, severely weakening the continuity of centralized monitoring and the efficiency of wind farm operation and maintenance.

[0003] In summary, existing technologies suffer from technical problems such as redundant fallback paths, repeated decisions, multiple path reconstructions, and uncontrollable recovery delays when maintenance request paths fail in multi-level monitoring architectures, which further affect the efficiency of wind farm operation and maintenance. Summary of the Invention

[0004] The purpose of this application is to provide a hierarchical operation and maintenance management method and system for wind farms based on multi-level monitoring, in order to solve the technical problems in the prior art where the path recovery mechanism of step-by-step fallback is used when the operation and maintenance request path fails in the multi-level monitoring architecture, resulting in redundancy of fallback paths, repeated decisions, multiple path reconstructions and uncontrollable recovery delays, which further affect the operation and maintenance efficiency of wind farms.

[0005] To achieve the above objectives, this application provides a method and system for hierarchical operation and maintenance management of wind farms based on multi-level monitoring.

[0006] Firstly, this application provides a hierarchical operation and maintenance management method for wind farms based on multi-level monitoring. This method is implemented through a hierarchical operation and maintenance management system for wind farms based on multi-level monitoring. The method includes: establishing a multi-level monitoring transmission network based on a set of monitoring terminal nodes; transmitting initial path data from the multi-level monitoring transmission network according to the wind farm operation and maintenance calculation task; triggering a rollback decision mechanism when a fault is detected at the next path node; constructing a target reachable node set using the target node of the wind farm operation and maintenance calculation task as a reference; analyzing and filtering the target reachable node set to obtain a candidate rollback node set; evaluating the candidate rollback node set to obtain a rollback cost index for each candidate rollback node, including an optional path quantification index, an optional path jump count index, and an optional path fault probability index; identifying a first rollback node using the rollback cost index; reconstructing a first jump path using the first rollback node as the successor node; and re-executing the operation and maintenance data transmission of the wind farm operation and maintenance calculation task according to the first jump path.

[0007] Optionally, the target node serves as a terminal constraint node, wherein the target node for the wind farm operation and maintenance calculation task includes a central control center node, an operation and maintenance data aggregation node, or a business processing node; obtain multi-dimensional state vectors of each node in the multi-level monitoring and transmission network, including node availability state vectors, resource load state vectors, and control capability level state vectors; perform a reverse reachability search on the multi-level monitoring and transmission network with the target node as the starting node for the reverse search to obtain an initial reachable node set; apply multi-level monitoring constraints to the initial reachable node set to obtain a target reachable node set.

[0008] Optionally, the multi-level monitoring constraints include control capability constraints characterizing the node's ability to carry operation and maintenance tasks, resource carrying constraints characterizing the node's load compatibility size, and topology level constraints restricting cross-topology levels; multi-level monitoring constraint analysis is performed on the initial set of reachable nodes, and the target set of reachable nodes is obtained by filtering those that meet the multi-level monitoring constraints.

[0009] Optionally, an empty set is initialized as the initial reachable node set, and the target node is added to the initial reachable node set as the starting node for reverse search. A breadth-first search is performed on the upstream nodes connected to the target node according to the multi-level monitoring and transmission network to obtain a breadth-first search node set. Reverse reachability search is continued according to the breadth-first search node set until the search node returns empty, at which point the initial reachable node set is output.

[0010] Optionally, an initial node set for the initial path is established; a preset jump step size is configured, and a neighborhood search is performed on the multi-level monitoring transmission network according to the preset jump step size to obtain a neighborhood initial node set including the initial node set; the intersection of the target reachable node set and the neighborhood initial node set is selected as a candidate fallback node set for output.

[0011] Optionally, the preset jump step size is used to control the neighborhood search depth of the multi-level monitoring transmission network. The method for calculating the preset jump step size includes: calculating the jump step size interval of the multi-level monitoring transmission network, and rounding up the average jump step size of the jump step size interval to obtain the preset jump step size.

[0012] Optionally, the task priority and data requirement type of the wind farm operation and maintenance calculation task are obtained; a first step length adjustment factor and a second step length adjustment factor are generated according to the task priority and data requirement type; the preset jump step length obtained by rounding up is optimized according to the first step length adjustment factor and the second step length adjustment factor, and the optimized preset jump step length is output.

[0013] Optionally, a set of optional paths for each candidate fallback node is obtained; wherein, the optional path quantification index is obtained by weighted calculation of the path length and node load change rate of the optional path set, the optional path jump count index is the average jump count index of all paths in the optional path set, and the optional path failure probability index is obtained by weighted fusion calculation of the probability accumulation of the historical failure probability of nodes and the historical failure probability of links of all paths in the optional path set.

[0014] Optionally, a set of optional paths with the first fallback node as the successor node is obtained, wherein the first fallback node is the candidate fallback node with the lowest fallback cost index in the candidate fallback node set; the set of optional paths is sorted according to the fallback cost index to obtain the first jump path.

[0015] Secondly, this application also provides a multi-level monitoring-based wind farm hierarchical operation and maintenance management system for executing the multi-level monitoring-based wind farm hierarchical operation and maintenance management method as described in the first aspect. The multi-level monitoring-based wind farm hierarchical operation and maintenance management system includes: a monitoring transmission network establishment module for establishing a multi-level monitoring transmission network based on a set of monitoring terminal nodes; an initial path transmission module for performing initial path transmission from the multi-level monitoring transmission network according to the wind farm operation and maintenance calculation task, triggering a rollback decision mechanism when a fault is detected in the next path node; and a rollback decision module for using the rollback decision mechanism based on the wind farm operation and maintenance calculation task. Using the target node as a reference, a set of target reachable nodes is constructed; a node filtering module is used to analyze and filter according to the target reachable node set to obtain a set of candidate rollback nodes; a rollback cost evaluation module is used to evaluate the set of candidate rollback nodes to obtain a rollback cost index for each candidate rollback node, including an optional path quantification index, an optional path jump count index, and an optional path failure probability index; a data retransmission module is used to identify the first rollback node using the rollback cost index, reconstruct the first jump path with the first rollback node as the successor node, and re-execute the operation and maintenance data transmission of the wind farm operation and maintenance calculation task according to the first jump path.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages: A multi-level monitoring and transmission network is established based on a set of monitoring terminal nodes; initial path transmission is performed from the multi-level monitoring and transmission network according to the wind farm operation and maintenance calculation task; a rollback decision mechanism is triggered when a fault is detected in the next path node; the rollback decision mechanism constructs a target reachable node set with the target node of the wind farm operation and maintenance calculation task as a reference; a candidate rollback node set is obtained by analyzing and filtering according to the target reachable node set; the rollback cost index of each candidate rollback node is obtained by evaluating the candidate rollback node set, including an optional path quantification index, an optional path jump count index, and an optional path fault probability index; the first rollback node is identified using the rollback cost index; a first jump path is reconstructed using the first rollback node as the successor node; and the operation and maintenance data transmission of the wind farm operation and maintenance calculation task is re-executed according to the first jump path. In other words, by establishing a multi-level monitoring and transmission network, when a failure of the next node is detected, a rollback decision mechanism is triggered. Taking the final target node of the task as a reference, a set of reachable nodes is constructed in reverse. Through evaluation of the candidate rollback node set and rollback cost indicators, the first rollback node with the best comprehensive cost is selected from multiple potential rollback points. A new path to the target is reconstructed, and data transmission continues along this path, realizing rapid self-healing of the global path and improving the reliability and operational efficiency of the entire wind farm hierarchical operation and maintenance system.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the hierarchical operation and maintenance management method for wind farms based on multi-level monitoring, as described in this application.

[0020] Figure 2 This is a schematic diagram of the structure of the wind farm hierarchical operation and maintenance management system based on multi-level monitoring in this application.

[0021] Figure labeling: 11 Monitoring transmission network establishment module, 12 Initial path transmission module, 13 Backoff decision module, 14 Node selection module, 15 Backoff cost evaluation module, 16 Data retransmission module. Detailed Implementation

[0022] This application provides a hierarchical operation and maintenance management method and system for wind farms based on multi-level monitoring. It addresses the technical problems in existing technologies where the path recovery mechanism, which uses a step-by-step fallback approach when an operation and maintenance request path fails in a multi-level monitoring architecture, leads to redundant fallback paths, repeated decisions, multiple path reconstructions, and uncontrollable recovery delays, further impacting the efficiency of wind farm operation and maintenance. By establishing a multi-level monitoring transmission network, when a failure of the next node is detected, a fallback decision mechanism is triggered. Using the final target node as a reference, a set of reachable nodes is constructed backwards. Through evaluation of the candidate fallback node set and fallback cost indicators, a first fallback node with the optimal overall cost is selected from multiple potential fallback points. A new path to the target is reconstructed, and data transmission continues along this path, achieving rapid self-healing of the global path and improving the reliability and operational efficiency of the entire wind farm hierarchical operation and maintenance system.

[0023] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0024] Example 1, please refer to the appendix. Figure 1 This application provides a hierarchical operation and maintenance management method for wind farms based on multi-level monitoring. The method is applied to a hierarchical operation and maintenance management system for wind farms based on multi-level monitoring. The specific steps of the method are as follows: Establish a multi-level monitoring transmission network based on a set of monitoring terminal nodes.

[0025] Specifically, after all wind farm monitoring terminal devices are powered on and complete self-tests, they acquire a pre-configured topology description file, which defines the identity information, hierarchical level, parent node, and maximum allowed communication hops for each node. Based on the topology description file, the neighbor discovery process is initiated level by level, starting with the central control layer node as the highest level. Nodes at each level periodically broadcast heartbeat probe messages, containing their own node identifier, current hierarchical level, and available resource status. Upon receiving a probe message, nodes at the same or adjacent levels record the sender's information and reply with an acknowledgment message, thereby establishing a bidirectional neighbor relationship table between nodes.

[0026] Connectivity verification is performed on each node. For edge-layer sensor nodes, such as a vibration sensor located in a wind turbine nacelle, an attempt is made to connect to its pre-defined site-level aggregation node, such as the wind turbine's edge computing gateway. If the connection is successful, the gateway node records the sensor as a direct child node; if the connection fails, an attempt is made to connect to the gateway node of an adjacent wind turbine according to the configured backup path. Simultaneously, site-level nodes periodically summarize their own status and the status of all their child nodes, reporting this information to the central control layer node via fiber optic or 5G network. The central control layer node collects the reported information from all sites, forming a complete multi-level monitoring transmission network topology. An initial health assessment is performed on the multi-level monitoring transmission network topology. Each node calculates the shortest path from itself to its assigned central control layer target node and stores this path information in its local routing table. Each path is marked with the sequence of intermediate nodes, the estimated transmission latency of each link, and the historical availability rate. For isolated nodes that cannot connect to any upper-level nodes, they are marked as faulty or offline, and maintenance personnel are prompted to check them in the monitoring interface. After completing neighbor discovery, connectivity verification, and initial path calculation for all the aforementioned nodes, the multi-level monitoring transmission network is declared successfully established and enters a standby state, awaiting the receipt of wind farm operation and maintenance calculation tasks.

[0027] The multi-level monitoring transmission network is a communication network organized according to the hierarchical structure of wind farm operation and maintenance management, namely the edge layer, the site layer, and the central control layer. Nodes in the multi-level monitoring transmission network have clearly defined hierarchical affiliations; upper-level nodes can manage lower-level nodes, and nodes at the same level can communicate as needed. Data transmission follows hierarchical constraints but allows limited cross-level jumps in case of failure. The multi-level monitoring transmission network is not a pre-fixed topology but a logical transmission network dynamically composed based on currently active and available monitoring terminal nodes. Its connection relationships are determined by the physical location of the nodes, signal coverage, hierarchical relationships, and preset communication rules.

[0028] The initial path transmission is performed from the multi-level monitoring and transmission network according to the wind farm operation and maintenance calculation task. When a failure is detected in the next path node, a rollback decision mechanism is triggered.

[0029] Specifically, when receiving wind farm operation and maintenance calculation tasks, these tasks are typically encapsulated into a data message. This message explicitly defines the source node identifier, target node identifier, task priority, and data payload. Each wind farm operation and maintenance calculation task has clearly defined source and target nodes. The source node queries the primary path field of its local routing table to obtain a pre-calculated initial path node sequence.

[0030] Following the initial path sequence, data packets are sent to the first next-path node in the sequence. The sending method uses either User Datagram Protocol (UDP) with acknowledgment or Simple Repeat Protocol (STP). That is, after the source node sends a packet, it starts a timer to wait for an acknowledgment packet from the next-path node. The timer's waiting duration is dynamically set based on the historical round-trip time of the link, typically three times the average round-trip time. For example, if the historical average time is 20ms, the timer is set to 60ms. During the waiting period, the source node concurrently monitors the health status of the next-path node. This is done by listening to the node's periodically sent heartbeat packets. Normal nodes broadcast a heartbeat every second, containing information such as the node's current load and temperature. If the source node neither receives a data acknowledgment packet nor a new heartbeat packet before the timer expires, or if the received heartbeat packet explicitly indicates that the node is overloaded or faulty, the source node determines that the next-path node has failed. Specific manifestations of a failure include node power failure, unresponsive network ports, heartbeat timeout, failure to receive acknowledgment after data packet transmission, and the node reporting resource exhaustion.

[0031] Once a fault is detected, the source node immediately suspends the current data transmission attempt and releases the cached resources used for this transmission. It then generates a fault event record, including the timestamp of the fault occurrence, the fault node identifier, the task identifier of the current task, and the task priority. Simultaneously, the source node sends back the context of the current task, including the source node, target node, task priority, data payload, and the initial paths already attempted, and enters a waiting state.

[0032] To prevent false positives, a fault-tolerant counter is typically set up. If no acknowledgment is received after three consecutive transmissions, but the heartbeat is still being received normally, a backoff decision is not immediately triggered. Instead, the data is attempted to be sent via the backup next path. Only when the backup path also fails, or the heartbeat is completely lost, is the node failure confirmed and a backoff decision triggered, thus avoiding unnecessary path switching due to brief network jitter.

[0033] Upon detecting a failure in the next path node, instead of simply retrying by stepping back to the previous node, a complete decision-making process is triggered. This process involves searching for reachable nodes in reverse order with the target node as a reference, filtering candidate backoff nodes, evaluating backoff cost metrics, and finally determining the optimal backoff node and reconstructing the jump path in one go.

[0034] The rollback decision mechanism uses the target node of the wind farm operation and maintenance calculation task as a reference to construct a set of target reachable nodes.

[0035] Furthermore, this application also includes the following steps: the target node serves as a terminal constraint node, wherein the target node of the wind farm operation and maintenance calculation task includes a central control center node, an operation and maintenance data aggregation node, or a business processing node; obtain multi-dimensional state vectors of each node in the multi-level monitoring and transmission network, including node availability state vectors, resource load state vectors, and control capability level state vectors; perform a reverse reachability search on the multi-level monitoring and transmission network with the target node as the starting node for the reverse search, to obtain an initial reachable node set; apply multi-level monitoring constraints to the initial reachable node set to obtain a target reachable node set.

[0036] Furthermore, this application also includes the following steps: the multi-level monitoring constraints include control capability constraints characterizing the node's ability to bear operation and maintenance tasks, resource bearing constraints characterizing the node's load compatibility size, and topology level constraints restricting cross-topology levels; multi-level monitoring constraint analysis is performed on the initial set of reachable nodes, and the target set of reachable nodes is obtained by screening those that meet the multi-level monitoring constraints.

[0037] Furthermore, this application also includes the following steps: initializing an empty set as an initial reachable node set, adding the target node as the starting node for reverse search to the initial reachable node set; performing a breadth-first search on the upstream nodes connected to the target node according to the multi-level monitoring transmission network to obtain a breadth-first search node set; continuing to perform a reverse reachability search according to the breadth-first search node set until the search node returns empty and outputting the initial reachable node set.

[0038] Specifically, the core of the rollback decision mechanism is to find all nodes that can reach the target node and satisfy multi-level monitoring constraints, providing a limited but effective search space for subsequent screening of candidate rollback nodes. When the rollback decision mechanism is triggered, the target node identifier is first extracted from the operation and maintenance computing task. The target node is the node that the wind farm operation and maintenance computing task ultimately needs to reach; depending on the task type, it can be the central control center server, the site data aggregation server, or a dedicated business processing server. The target node is set as the terminal constraint node, i.e., the endpoint of all new paths.

[0039] The system retrieves the multi-dimensional state vector of each node in the current multi-level monitoring and transmission network from the network management module. This vector is dynamically updated via heartbeat messages and periodic reporting and stored in a centralized topology database. The multi-dimensional state vector includes node availability state vector, resource load state vector, and control capability level state vector. The node availability state vector records whether each node is online and fault-free. The resource load state vector records the node's current CPU utilization, memory utilization, and remaining network bandwidth. The control capability level state vector records the node's level and its capability label, with the level represented by numbers, such as 0 for the central control layer, 1 for the field station layer, and 2 for the edge layer.

[0040] Create an empty set of nodes, i.e., the initial reachable node set. Add the target node itself as the starting node for the reverse search to the initial reachable node set. The target node is included because it is inherently reachable from itself, i.e., a zero-hop path. When filtering candidate fallback nodes, if a node upstream of the failed node happens to be the target node itself, it can be directly sent to.

[0041] Create a first-in, first-out queue named the "queue for nodes to be searched". Add the target node to the queue. Simultaneously, create a table marking visited nodes to record which nodes have already been searched, avoiding duplicate processing. Initially, mark the target node as visited. Query the topology database of the multi-level monitoring transmission network to find all nodes that can directly send data to the target node; these nodes are called the target node's upstream nodes. Determine if a directed edge exists from a node X to the target node. Record all these upstream nodes and add them to a newly created set named the "breadth-first search node set". This set only includes upstream nodes that are one hop away from the target node (i.e., directly connected), excluding nodes with more than one hop. Also add these upstream nodes to the initial reachable node set, as the target node can obviously be reached in one step from these nodes. At this point, the initial reachable node set contains the target node and all its direct upstream nodes.

[0042] The breadth-first search node set is used as the current node set to be searched. A new empty set is created, named the next-level upstream node set, to store newly found upstream nodes in this round. For each node in the current node set to be searched (i.e., the current node), the topology database is queried to find all upstream nodes that can directly send data to the current node (i.e., the current node's upstream nodes). For each found upstream node, it is checked whether it is already in the initial reachable node set (i.e., whether it has been visited). If not, the upstream node is added to the initial reachable node set and simultaneously added to the next-level upstream node set. After traversing all current nodes, the next-level upstream node set is checked to see if it is empty. If it is empty, it means that no new upstream node can be found, and the search terminates. If it is not empty, the next-level upstream node set is used as the new current node set to be searched, and the above loop is repeated, i.e., continuing to expand one level outward. When the search terminates, the initial reachable node set is output, which includes the target node itself and all nodes that can eventually reach the target node through forward data transmission (i.e., all nodes reachable from the topology). The entire search process expands outward layer by layer: the first layer is the direct upstream of the target node, the second layer is the upstream of these upstream nodes, and so on, until the edge of the network, to obtain the initial set of reachable nodes.

[0043] The initial set of reachable nodes may include some nodes that, while topologically reachable, are actually unable to handle the operational tasks. Multi-level monitoring constraints are applied to this initial set of reachable nodes, checking control capability constraints, resource capacity constraints, and topology layer constraints one by one. Control capability constraints check whether a node has the capacity to handle the current operational task. Resource capacity constraints compare the resource requirements of the current operational task with the node's current resource load. Topology layer constraints check whether the node's layer in the multi-level monitoring architecture meets the cross-level transmission rules. For example, first, the control capability constraint is checked: if the task is forwarding massive amounts of real-time data, the node must have a forwarding capability label; if the task is performing edge computing, the node must have a computation capability label. Nodes lacking the corresponding capabilities are removed. Next, resource capacity constraints are applied: the task requires an additional 5% of the CPU and 10 megabytes of memory. If the node's current CPU utilization plus 5% does not exceed the node's maximum allowed threshold (usually set to 85%), and the remaining memory is greater than 10 megabytes, then the constraint is met; otherwise, the node is removed. Finally, the topology hierarchy constraints are checked. Edge layer nodes can only send messages directly to station layer nodes or specific relay nodes at the same level, not directly to control layer nodes. Based on the node's hierarchy vector and preset rules, it is determined whether the node is allowed to appear in the fallback path, i.e., whether it is allowed to serve as a fallback node. Nodes that do not meet the hierarchy specifications will be eliminated. After filtering by the above three constraints, all remaining nodes in the initial reachable node set constitute the target reachable node set. Each node in this set can serve as a candidate for subsequent fallback nodes because they can reach the target node, have the ability to undertake the task, sufficient resources, and compliant hierarchy relationships.

[0044] The final set of nodes is output to the next stage of the rollback decision mechanism, namely the stage of analyzing and screening to obtain the candidate rollback node set. The size of this set is usually much smaller than the total number of nodes in the entire network, which greatly reduces the search space for subsequent cost evaluation while ensuring the effectiveness of the nodes in the set.

[0045] By performing a reverse search with the target node as a reference, it ensures that each node in the set has at least one valid path to the target node, fundamentally solving the problem of blindly backing up to nodes that cannot reach the target in step-by-step backtracking. By eliminating invalid nodes through constraints, the reverse search itself is limited to including only nodes that can reach the target; compared to a full network search, the decision space changes from unconstrained to constrained. Control capability constraints ensure that nodes have the basic functions to perform tasks; resource capacity constraints prevent tasks from being assigned to overloaded nodes, preventing secondary failures; and topology hierarchy constraints ensure that paths conform to the hierarchical specifications of wind farm monitoring, avoiding management chaos caused by out-of-order jumps.

[0046] The candidate fallback node set is obtained by analyzing and filtering the set of reachable nodes.

[0047] Furthermore, this application also includes the following steps: establishing an initial node set for the initial path; configuring a preset jump step size, performing a neighborhood search on the multi-level monitoring transmission network according to the preset jump step size, and obtaining a neighborhood initial node set including the initial node set; and filtering the intersection of the target reachable node set and the neighborhood initial node set as a candidate fallback node set for output.

[0048] Furthermore, this application also includes the following steps: the preset jump step size is used to control the neighborhood search depth of the multi-level monitoring transmission network, and the method for calculating the preset jump step size includes: calculating the jump step size interval of the multi-level monitoring transmission network, and rounding up the average jump step size of the jump step size interval to obtain the preset jump step size.

[0049] Furthermore, this application also includes the following steps: obtaining the task priority and data requirement type of the wind farm operation and maintenance calculation task; generating a first step length adjustment factor and a second step length adjustment factor according to the task priority and data requirement type; optimizing the preset jump step size obtained by rounding up according to the first step length adjustment factor and the second step length adjustment factor, and outputting the optimized preset jump step size.

[0050] Specifically, when the rollback decision mechanism is triggered, the initial transmission path used by the current operation and maintenance computing task is retrieved from local storage. All nodes within this path are extracted to form an initial node set, which includes all nodes traversed by the initial path, including the source node, all intermediate forwarding nodes, and the target node. Even if some nodes may have failed due to faults, they are still recorded in the initial path. If the initial path contains a faulty node, that node is also part of the initial node set because it is part of the original path. However, in subsequent neighborhood searches, since faulty nodes may not be responsive, they are usually only treated as location reference points and no data is actually sent to them.

[0051] A preset jump step size is configured to control the neighborhood search depth of the multi-level monitoring transmission network, preventing an excessively large search range from increasing computational overhead, or an excessively small range from missing potential fallback nodes. The preset jump step size represents the maximum number of hops that can be explored forward or backward from the initial node during neighborhood search. The topology of the entire multi-level monitoring transmission network is analyzed, the shortest path hop count between all node pairs is calculated, and the minimum and maximum values ​​are taken to obtain the jump step size interval. For example, if the minimum hop count is 2 and the maximum hop count is 8, the midpoint of this interval is rounded up as the base step size. The specific formula is: base step size = ceil((minimum hop count + maximum hop count) / 2), resulting in a balanced initial step size that covers a certain neighborhood range without being too large.

[0052] The task priority and data requirement type are read from the operation and maintenance computing tasks. Task priority represents the urgency or importance of the current operation and maintenance computing task, usually expressed numerically. Data requirement type describes the data transmission quality requirements of the operation and maintenance computing task, including real-time sensitive, bandwidth sensitive, and reliability sensitive types. Based on preset mapping rules, a first-step adjustment factor and a second-step adjustment factor are generated. The first-step adjustment factor is a coefficient generated based on task priority, used to scale the baseline step size. Higher priority typically results in a factor less than 1, reducing the step size and prioritizing neighboring nodes; lower priority results in a factor greater than 1, allowing for a larger step size and expanding the search range. The second-step adjustment factor is a coefficient generated based on data requirement type, used to further fine-tune the step size. For example, for highly real-time data, the factor is less than 1; for high-bandwidth data, the factor may be slightly greater than 1 to allow for more flexible path selection. The optimized preset jump step size = preset jump step size × first-step adjustment factor × second-step adjustment factor, and the result is rounded to the nearest integer, ensuring the final step size is at least 1.

[0053] Using each node in the initial node set as a center point, a neighborhood search is performed. For each center node, the search expands outward along the links in the network topology, recording all nodes within a preset hop distance of that node. A breadth-first search is used, starting from the center node and expanding outward step by step until the preset hop distance is reached. After searching all center nodes, all found nodes are merged into a single set, and duplicate nodes are removed, forming the initial neighborhood node set. This set represents all nodes within a certain range around the initial path, and these nodes are topologically close to the initial path.

[0054] Calculate the intersection of the target reachable node set and the initial neighborhood node set. That is, find those nodes that are both in the target reachable node set and within the vicinity of the initial path, and use them as a candidate backtracking node set. If the candidate backtracking node set is empty, it means there are no nodes near the initial path that can reach the target node. In this case, increase the preset jump step size and re-search the neighborhood. Ensure that each node in the target reachable node set satisfies two conditions: it can eventually reach the target node and it is located within the vicinity of the initial path.

[0055] For example, assume the initial transmission path is: edge gateway node G02 of wind turbine No. 2, relay node R03, site server S01, and central control center C01. A fault occurs at relay node R03, and a candidate fallback node set is selected. The initial path node sequence is G02, R03, S01, and C01. After deduplication, the initial node set is G02, R03, S01, and C01. At this point, R03 has failed but is still in the set as a location reference. The wind farm multi-level monitoring transmission network has a total of 154 nodes. Calculations show that the minimum hop count for the shortest path between any two nodes is 1 (for directly connected nodes), and the maximum is 6. Therefore, the jump step size interval is [1, 6], and the basic step size = ceil((1+6) / 2) = 4, where ceil is a floor function that returns the smallest integer greater than or equal to a given value. The current task attributes are read: the task is to upload vibration data from wind turbine No. 2, with a priority of level two, a data requirement type of "regular," and medium real-time requirements with low bandwidth requirements. According to the preset mapping table: the first step length adjustment factor for level two priority is 0.8; the second step length adjustment factor for regular data is 1.0. The optimized step size = 4 × 0.8 × 1.0 = 3.2, rounded to 3. Therefore, the preset jump step size = 3. A neighborhood search is performed with each node in the initial node set as the center, with a hop count not exceeding 3. Centered on G02, nodes within 1 hop include G02 itself and its neighbors R03, G01, and G02. Nodes within 2 hops include R02, R04, and some sensors. Nodes within 3 hops further extend to the more distant R01, R05 (which is faulty), and more sensors. Statistics show that approximately 28 nodes are obtained within a 3-hop range starting from G02. Centered on R03 (though the topology is still usable despite the failure), the first hop includes R03 itself, and its neighbors are G02, G03, G04, and S01, since R03 is upstream of S01. The second hop includes G01, G05, S02, as well as R02, R04, etc. The third hop extends to more gateways and sensors. Approximately 35 nodes are identified, with significant overlap with G02's neighborhood. Centered on S01, the first hop includes S01 itself, and its neighbors are R01 to R10, as well as C01 and S02. The second hop includes all downstream gateways of R nodes, and the downstream of C01. C01 is the control layer, and its downstream is S01 itself, so it is not extended further. Approximately 45 nodes are identified. Centered on C01, the first hop includes C01 itself, and its neighbors are S01 and S02. Distance 2 hops includes all upstream nodes of S01 and S02, i.e., all R nodes and the gateway. Distance 3 hops includes all sensors under all R nodes. The total number of nodes is approximately 95, almost the entire reachable network. After merging all found nodes and removing duplicates, due to significant overlap among the four neighborhoods, the final initial set of neighborhood nodes is approximately 98 nodes, close to the entire reachable network.The specific nodes include C01, S01, S02, all relay nodes R01 to R15, most gateways G01 to G20 (except R05), and approximately 60 sensors. The target reachable node set is 95 nodes, including C01, S01, S02, 14 relays, 20 gateways, and 58 sensors. The initial neighborhood node set is 98 nodes. The intersection of the two is almost the target reachable node set itself, because most nodes in the target reachable node set are within 3 hops of the initial path. The few nodes not in the neighborhood are those sensors that are more than 3 hops away from the initial path. After precise calculation, the intersection size is 92 nodes, specifically including C01, S01, S02, all 14 relays, 20 gateways, and 55 sensors. Therefore, the candidate fallback node set includes C01, S01, S02, R01, R02, R03, R04, R06...R15, G01...G20, and 55 sensor nodes, totaling 92 nodes. If the task priority is higher, the step size adjustment factor is 0.6, then the preset jump step size = 4 × 0.6 × 1.0 = 2.4, rounded down to 2. At this time, the neighborhood search range is reduced, and the initial neighborhood node set obtained is only about 50 nodes. The intersection will be significantly reduced, there will be fewer candidate nodes, and the decision will be faster.

[0056] The candidate fallback node set is a further selection of nodes that are close to the initial path from the target reachable node set. It must meet two conditions at the same time: reachable target node and located near the initial path. This prevents the selection of fallback nodes that are reachable but too far away, thereby avoiding additional latency and resource consumption caused by long-distance fallback.

[0057] The set of candidate fallback nodes is evaluated to obtain fallback cost indicators for each candidate fallback node, including optional path quantification indicators, optional path jump count indicators, and optional path failure probability indicators.

[0058] Furthermore, this application also includes the following steps: obtaining a set of optional paths for each candidate fallback node; wherein, the optional path quantification index is obtained by weighted calculation of the path length and node load change rate of the optional path set, the optional path jump count index is the average jump count index of all paths in the optional path set, and the optional path failure probability index is obtained by weighted fusion calculation of the probability accumulation of the node historical failure probability and the link historical failure probability of all paths in the optional path set.

[0059] Specifically, for each candidate fallback node in the candidate fallback node set, all paths from that node to the target node are identified. The fully connected graph of the current network is retrieved from the topology database. A constrained depth-first search is run, starting from the current candidate node and ending at the target node, to find the top K shortest paths. The value of K is set according to the network size and real-time requirements, such as K=5. During the enumeration process, hierarchical constraints of multi-level monitoring need to be applied, such as prohibiting edge layer nodes from directly connecting to control layer nodes and avoiding loops. For each path found, the node sequence, link sequence, total hop count, and estimated latency of each link are recorded. For the current candidate node, a set of optional paths is obtained, containing K paths from that node to the target node. The set of optional paths includes all possible paths from each candidate fallback node to the target node, and must strictly adhere to the topology of the multi-level monitoring transmission network and satisfy basic hierarchical constraints.

[0060] For each path in the set of optional paths, calculate the path length and node load change rate. Obtain the quantified value of the path through a weighted summation. Finally, average the quantified values ​​of all paths as the quantified index for the candidate node's optional path. Path length is typically defined as the number of hops on the path. The node load change rate is evaluated by calculating the proportion of additional resource load that each node along the path would need to bear relative to its remaining capacity if the path were used as a new transmission path. For each node on the path, obtain its current resource load rate. Based on the resource requirements of the operation and maintenance computation task, calculate the predicted load rate after the task is added. If the predicted load rate exceeds the node's maximum safety threshold, such as 85%, the load change rate is counted as infinity, indicating unavailability; otherwise, the load change rate = resource requirement / node remaining capacity, where node remaining capacity = maximum safety threshold - current load rate. For example, if the current load is 70% and the maximum safety threshold is 85%, then the remaining capacity is 15%, the task requires 5%, and the load change rate = 0.333. The average load change rate of all nodes on the path is taken as the load change rate component for that path.

[0061] The path length and node load change rate are weighted and summed. The path quantification value = α × path length + β × node load change rate, where α and β are preset weighting coefficients, such as α = 0.4 and β = 0.6, to balance the influence of hop count and load. The average of the path quantification values ​​of all possible paths to the candidate node is taken to obtain the candidate node's possible path quantification index.

[0062] For each path in the set of optional paths, record its hop count, which is the number of nodes on the path minus 1. Calculate the arithmetic mean of the hop counts of all paths to obtain the hop count index of the optional paths for that candidate node. This index reflects the average number of hops required to reach the target node from that node. The fewer the hop count, the lower the transmission latency and the higher the reliability.

[0063] For each path in the set of optional paths, the historical failure probabilities of each node and each link on the path are considered comprehensively. The failure probability of the entire path is calculated using a probability accumulation method. First, the survival probability of each node and each link on the path is calculated separately, then multiplied to obtain the survival probability of the entire path. Finally, the failure probability is obtained by subtracting the survival probability from 1. Then, the failure probabilities of all optional paths are weighted and fused, such as by weighted averaging, to obtain the optional path failure probability. Node historical failure probability The historical link failure probability is used to obtain the failure risk index of the candidate fallback node. The lower the index, the more stable the path. Here, 'a' and 'b' are the weights of the node's historical failure probability and the link's historical failure probability, respectively, such as a=0.5 and b=0.5.

[0064] For each candidate fallback node, its optional path quantification metric, optional path hop count metric, and optional path failure probability metric are stored to obtain the fallback cost metric. The optional path quantification metric considers both path length and node load change rate, thus ensuring transmission efficiency while preventing the initiation of new node failures. The hop count metric intuitively provides the average number of hops required; fewer hops generally result in lower cumulative latency and failure probability, helping to select nodes with closer physical distances and better topologies. By fusing the historical failure probabilities of nodes and links, the long-term stability of each path can be predicted, avoiding the selection of paths with few hops but frequent failures, thereby improving the reliability of operational data transmission.

[0065] The first rollback node is identified using the rollback cost index, and the first jump path is reconstructed using the first rollback node as the successor node. The operation and maintenance data transmission of the wind farm operation and maintenance calculation task is re-executed according to the first jump path.

[0066] Furthermore, this application also includes the following steps: obtaining a set of optional paths with the first fallback node as the successor node, wherein the first fallback node is the candidate fallback node with the lowest fallback cost index in the candidate fallback node set; sorting the set of optional paths according to the fallback cost index to obtain a first jump path.

[0067] Specifically, from the candidate fallback node set, the pre-calculated fallback cost metrics for each node are extracted, including the optional path quantification metric, the number of jumps metric, and the failure probability metric. These three metrics are then ranked, for example, first comparing the failure probability metric, then the number of jumps metric, and finally the quantification metric. Based on this ranking, weights are assigned to the optional path quantification metric, the number of jumps metric, and the failure probability metric, such as 0.5, 0.3, and 0.2. The optional path quantification metric... Optional path quantification metrics weight + jump count metrics Jump count metric weight + failure probability metric The overall cost is calculated using the weighted failure probability index. The node with the lowest overall cost is selected as the first fallback node. If multiple nodes have the same cost, they are selected according to their node numbers.

[0068] Starting from the first fallback node and ending at the target node, a restricted path search is performed again to obtain a set of optional paths from the first fallback node to the target node. Since the network topology may change after a failure, this re-search ensures the path information is up-to-date. Each path records the node sequence, link sequence, hop count, and the real-time status of each node and link. These optional paths are added to a set called the optional path set with the first fallback node as the successor node. If this set is empty, meaning there is no path from the first fallback node to the target node, it indicates that the fallback node is actually unreachable. In this case, it is necessary to fall back to the second-best node in the candidate fallback node set and re-execute.

[0069] For each path in the set of optional paths, a cost evaluation method is used to recalculate the comprehensive cost of each path. The paths are sorted in ascending order of comprehensive cost, and the path with the lowest cost is the optimal path. This optimal path is marked as the first jump path, including the node currently holding the data (i.e., the upstream node of the faulty node), and then directly points to the first fallback node. The path then proceeds through the intermediate nodes sequentially according to the optimal path, finally reaching the target node. Therefore, the number of hops in the first jump path may be slightly higher than the original path because it needs to bypass the faulty node, but through the optimization of this scheme, the increase is kept to a minimum.

[0070] Upon obtaining the first jump path, the data transmission recovery process is immediately initiated. The current node repackages the data packet of the operation and maintenance computing task, adding the node sequence of the new path to the packet header. The current node sends the data packet to the first node on the first jump path, i.e., the first fallback node. A reliable transmission mechanism, such as retransmission with acknowledgment, is used during transmission, and a timeout timer is set. After receiving the packet, the first fallback node parses the subsequent path information and forwards the packet to the next node on the path, and so on, until the packet reaches the target node. After receiving the packet, the target node returns an end-to-end acknowledgment packet, which is then transmitted back to the current node along the original path. Upon receiving the acknowledgment, the current node marks the operation and maintenance computing task as successfully transmitted. If a node failure occurs again during transmission, the fallback decision mechanism of this invention is triggered again, and a complete fault self-healing process is repeated.

[0071] To ensure no data loss, the current node typically caches data packets before sending them, releasing the cache only upon receiving acknowledgment from the target node. If no acknowledgment is received within a timeout period, the packet is retransmitted; if this also fails, a rollback decision is triggered again. The data from the initially interrupted wind farm operation and maintenance calculation task is then retransmitted along a new path, ensuring no data loss or duplication, and arriving at the target node in order.

[0072] For example, the initial path is: G02→R03→S01→C01, where R03 fails. After evaluation, the candidate fallback node set is determined to be the first fallback node because it has the lowest overall cost, with the three indicators being a quantification score of 2.081, a jump count of 1.5, and a failure probability of 0.325%. The overall cost is compared among the three nodes R02, R04, and S01 in the candidate fallback node set. The overall cost is calculated using a normalized weighted method, assuming equal weights, by normalizing each indicator and summing them. Using S01 as the benchmark (quantification 2.081, jump count 1.5, failure probability 0.325), R02 (quantification 3.125, jump count 2.5, failure probability 1.00), and R04 (quantification 3.258, jump count 2.5, failure probability 1.375), the calculated overall score for S01 is the lowest, therefore S01 is selected as the first fallback node. Starting from S01 and targeting C01, a new path search is performed. The current network status is that R03 is faulty, while the other nodes are normal. The search yields three valid paths: Path A is S01→C01 (hop count = 1), directly connected via fiber optic cable, with a latency of approximately 15ms; Path B is S01→S02→C01 (hop count = 2), with S02 being a backup site server, and a latency of approximately 30ms; Path C is S01→R02→S01→C01 (forming a loop, not allowed, excluded). Only Path A and Path B are actually valid paths. The backoff cost metrics for Path A and Path B are recalculated: Path A (1 hop) has an optional path quantification metric of 1.581, an optional path hop count metric of 1, and an optional path failure probability metric of 0.25%. The overall cost is slightly lower than that of Path B. Path B (2 hops) has an optional path quantification metric of 2.581, an optional path hop count metric of 2, and an optional path failure probability metric of approximately 0.4%. The overall cost is higher than path A. Path A is selected as the first hop path. The node currently caching data is G02 (edge ​​gateway of wind turbine No. 2), which was originally intended to be sent to R03 (which has failed). Now G02 obtains the first hop path, and the node sequence of path A is S01→C01. However, the complete path needs to start from G02, so the actual first hop path is G02→S01→C01. In fact, there is a direct link between G02 and S01, so the hop count is 2. G02 re-encapsulates the data packet, specifying the next hop as S01 in the packet header, and carrying the final destination C01. G02 sends the packet through its direct link with S01. The sending time is 10:00:10.055, 55ms after fault detection. S01 receives the packet at 10:00:10.065, parses it, and immediately forwards it to C01. C01 receives the packet at 10:00:10.080. After verifying data integrity, C01 returns an acknowledgment message to G02. G02 receives the acknowledgment at 10:00:10.105 and releases its buffer. The entire recovery process, from fault triggering to data arrival at the target, takes only 28 milliseconds.If there is no direct link between G02 and S01, it needs to be relayed through R02, with the path being G02→R02→S01→C01 (hop number 3). In this case, the recovery latency is approximately 50ms, which is still far superior to step-by-step rollback. The maintenance platform shows that the vibration data of wind turbine No. 2 has been successfully uploaded with a latency of 80ms. This fault and self-healing event are recorded, and the status of R03 is updated to "fault pending repair."

[0073] Unlike step-by-step fallback, which involves multiple retries and reconstructions, this approach directly jumps to the globally optimal first fallback node and selects the best sub-path, reducing recovery latency from an uncontrollable linear increase to a manageable, fixed range. Through pre-send caching and end-to-end acknowledgment mechanisms, it ensures that even if a new path fails again, data is not lost, triggering another fallback decision. By re-searching the latest topology when acquiring the set of available paths, it can handle changes in the state of other nodes after a failure, avoiding the use of outdated path information.

[0074] In summary, the wind farm hierarchical operation and maintenance management method based on multi-level monitoring provided in this application has the following technical effects: A multi-level monitoring transmission network is established based on a set of monitoring terminal nodes; initial path transmission is performed from the multi-level monitoring transmission network according to the wind farm operation and maintenance calculation task; a rollback decision mechanism is triggered when a fault is detected in the next path node; the rollback decision mechanism constructs a target reachable node set with the target node of the wind farm operation and maintenance calculation task as a reference; a candidate rollback node set is obtained by analyzing and filtering according to the target reachable node set; the rollback cost index of each candidate rollback node is evaluated, including a selectable path quantification index, a selectable path jump count index, and a selectable path fault probability index; the first rollback node is identified using the rollback cost index; a first jump path is reconstructed using the first rollback node as the successor node; and the operation and maintenance data transmission of the wind farm operation and maintenance calculation task is re-executed according to the first jump path. In other words, by establishing a multi-level monitoring and transmission network, when a failure of the next node is detected, a rollback decision mechanism is triggered. Taking the final target node of the task as a reference, a set of reachable nodes is constructed in reverse. Through evaluation of the candidate rollback node set and rollback cost indicators, the first rollback node with the best comprehensive cost is selected from multiple potential rollback points. A new path to the target is reconstructed, and data transmission continues along this path, realizing rapid self-healing of the global path and improving the reliability and operational efficiency of the entire wind farm hierarchical operation and maintenance system.

[0075] Example 2: Based on the same inventive concept as the multi-level monitoring-based wind farm hierarchical operation and maintenance management method in Example 1, this application also provides a multi-level monitoring-based wind farm hierarchical operation and maintenance management system. Please refer to the appendix. Figure 2The wind farm hierarchical operation and maintenance management system based on multi-level monitoring includes: a monitoring transmission network establishment module 11, used to establish a multi-level monitoring transmission network based on a set of monitoring terminal nodes; an initial path transmission module 12, used to transmit the initial path from the multi-level monitoring transmission network according to the wind farm operation and maintenance calculation task, and trigger a rollback decision mechanism when a fault is detected in the next path node; a rollback decision module 13, used by the rollback decision mechanism to construct a target reachable node set with the target node of the wind farm operation and maintenance calculation task as a reference; a node filtering module 14, used to analyze and filter according to the target reachable node set to obtain a candidate rollback node set; a rollback cost evaluation module 15, used to evaluate the candidate rollback node set to obtain a rollback cost index for each candidate rollback node, including an optional path quantification index, an optional path jump count index, and an optional path fault probability index; and a data retransmission module 16, used to identify the first rollback node using the rollback cost index, reconstruct the first jump path with the first rollback node as the successor node, and re-execute the operation and maintenance data transmission of the wind farm operation and maintenance calculation task according to the first jump path.

[0076] Furthermore, the rollback decision module 13 in the wind farm hierarchical operation and maintenance management system based on multi-level monitoring is also used for: using the target node as a terminal constraint node, wherein the target node of the wind farm operation and maintenance calculation task includes a central control center node, an operation and maintenance data aggregation node, or a business processing node; obtaining multi-dimensional state vectors of each node in the multi-level monitoring transmission network, including node availability state vector, resource load state vector, and control capability level state vector; performing a reverse reachability search on the multi-level monitoring transmission network with the target node as the starting node for the reverse search, to obtain an initial reachable node set; and applying multi-level monitoring constraints to the initial reachable node set to obtain a target reachable node set.

[0077] Furthermore, the rollback decision module 13 in the wind farm hierarchical operation and maintenance management system based on multi-level monitoring is also used for: the multi-level monitoring constraints include control capability constraints characterizing the node's ability to carry operation and maintenance tasks, resource carrying constraints characterizing the node's load compatibility size, and topology level constraints restricting cross-topology levels; performing multi-level monitoring constraint analysis on the initial set of reachable nodes, and filtering out the target set of reachable nodes that meet the multi-level monitoring constraints.

[0078] Furthermore, the rollback decision module 13 in the multi-level monitoring-based wind farm hierarchical operation and maintenance management system is also used for: initializing an empty set as an initial reachable node set, adding the target node as the starting node for reverse search to the initial reachable node set; performing a breadth-first search on the upstream nodes connected to the target node according to the multi-level monitoring transmission network to obtain a breadth-first search node set; and continuing to perform a reverse reachability search according to the breadth-first search node set until the search node returns empty and the initial reachable node set is output.

[0079] Furthermore, the node filtering module 14 in the multi-level monitoring-based wind farm hierarchical operation and maintenance management system is also used to: establish an initial node set for the initial path; configure a preset jump step size, perform a neighborhood search on the multi-level monitoring transmission network according to the preset jump step size, and obtain a neighborhood initial node set including the initial node set; and filter the intersection of the target reachable node set and the neighborhood initial node set as a candidate fallback node set for output.

[0080] Furthermore, the node screening module 14 in the wind farm hierarchical operation and maintenance management system based on multi-level monitoring is also used for: the preset jump step size is used to control the neighborhood search depth of the multi-level monitoring transmission network, and the method for calculating the preset jump step size includes: calculating the jump step size interval of the multi-level monitoring transmission network, and rounding up the average jump step size of the jump step size interval to obtain the preset jump step size.

[0081] Furthermore, the node filtering module 14 in the wind farm hierarchical operation and maintenance management system based on multi-level monitoring is also used to: obtain the task priority and data requirement type of the wind farm operation and maintenance calculation task; generate a first step length adjustment factor and a second step length adjustment factor according to the task priority and data requirement type; optimize the preset jump step length obtained by rounding up according to the first step length adjustment factor and the second step length adjustment factor, and output the optimized preset jump step length.

[0082] Furthermore, the rollback cost assessment module 15 in the wind farm hierarchical operation and maintenance management system based on multi-level monitoring is also used to: obtain the set of optional paths for each candidate rollback node; wherein, the optional path quantification index is obtained by weighted calculation of the path length and node load change rate of the optional path set, the optional path jump count index is the average jump count index of all paths in the optional path set, and the optional path failure probability index is obtained by cumulative weighted fusion calculation of the probability of historical node failure and historical link failure of all paths in the optional path set.

[0083] Furthermore, the data retransmission module 16 in the wind farm hierarchical operation and maintenance management system based on multi-level monitoring is also used to: obtain a set of optional paths with the first backoff node as the successor node, wherein the first backoff node is the candidate backoff node with the lowest backoff cost index in the candidate backoff node set; and sort the set of optional paths according to the backoff cost index to obtain a first jump path.

[0084] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The wind farm hierarchical operation and maintenance management method and specific examples based on multi-level monitoring in the aforementioned embodiment 1 are also applicable to the wind farm hierarchical operation and maintenance management system based on multi-level monitoring in this embodiment. Through the foregoing detailed description of the wind farm hierarchical operation and maintenance management method based on multi-level monitoring, those skilled in the art can clearly understand the wind farm hierarchical operation and maintenance management system based on multi-level monitoring in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0085] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0086] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A hierarchical operation and maintenance management method for wind farms based on multi-level monitoring, characterized in that, include: Establish a multi-level monitoring transmission network based on a set of monitoring terminal nodes; Based on the wind farm operation and maintenance calculation task, the initial path transmission is carried out from the multi-level monitoring and transmission network. When a failure is detected in the next path node, a fallback decision mechanism is triggered. The rollback decision mechanism uses the target node of the wind farm operation and maintenance calculation task as a reference to construct a set of target reachable nodes; A candidate fallback node set is obtained by analyzing and filtering the target reachable node set. The candidate fallback node set is evaluated to obtain fallback cost indicators for each candidate fallback node, including optional path quantification indicators, optional path hop count indicators, and optional path failure probability indicators. The optional path quantification indicators are obtained by weighted calculation of the path length and node load change rate of the optional path set. The optional path hop count indicators are the average hop count indicators of all paths in the optional path set. The optional path failure probability indicators are obtained by cumulative weighted fusion calculation of the historical failure probabilities of nodes and links of all paths in the optional path set. The first rollback node is identified using the rollback cost index, and the first jump path is reconstructed using the first rollback node as the successor node. The operation and maintenance data transmission of the wind farm operation and maintenance calculation task is re-executed according to the first jump path. The rollback decision mechanism, taking the target node of the wind farm operation and maintenance calculation task as a reference, constructs a set of target reachable nodes, and the method includes: The target node serves as the terminal constraint node, wherein the target node of the wind farm operation and maintenance calculation task includes the central control center node, the operation and maintenance data aggregation node, or the business processing node. Obtain multi-dimensional state vectors of each node in the multi-level monitoring and transmission network, including node availability state vector, resource load state vector, and control capability level state vector; Using the target node as the starting node for the reverse search, a reverse reachability search is performed on the multi-level monitoring and transmission network to obtain an initial set of reachable nodes. Multi-level monitoring constraints are applied to the initial set of reachable nodes to obtain the target set of reachable nodes.

2. The wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in claim 1, characterized in that, The multi-level monitoring constraints include control capability constraints that characterize the node's ability to carry operation and maintenance tasks, resource carrying constraints that characterize the node's load compatibility size, and topology level constraints that restrict cross-topology levels. A multi-level monitoring constraint analysis is performed on the initial set of reachable nodes, and the target set of reachable nodes is obtained by filtering those that meet the multi-level monitoring constraints.

3. The wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in claim 1, characterized in that, Using the target node as the starting node for the reverse search, a reverse reachability search is performed on the multi-level monitoring transmission network to obtain an initial set of reachable nodes. The method includes: Initialize an empty set as the initial reachable node set, and add the target node as the starting node for the reverse search to the initial reachable node set; Based on the multi-level monitoring and transmission network, a breadth-first search is performed on the upstream nodes connected to the target node to obtain a breadth-first search node set; The reverse reachability search continues based on the breadth-first search node set until the search node returns empty, at which point the initial reachable node set is output.

4. The wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in claim 1, characterized in that, The method for obtaining a candidate fallback node set by analyzing and filtering the target reachable node set includes: Establish the initial node set for the initial path; Configure a preset jump step size, and perform a neighborhood search on the multi-level monitoring transmission network according to the preset jump step size to obtain a neighborhood initial node set including the initial node set; The intersection of the target reachable node set and the initial neighborhood node set is selected as the candidate backtracking node set for output.

5. The wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in claim 4, characterized in that, The preset jump step size is used to control the neighborhood search depth of the multi-level monitoring transmission network. The method for calculating the preset jump step size includes: The preset jump step size is obtained by calculating the jump step size range of the multi-level monitoring transmission network and rounding up the average jump step size of the jump step size range.

6. The wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in claim 5, characterized in that, The preset jump step size is calculated by rounding up the average jump step size of the jump step size interval. The method further includes: Obtain the task priority and data requirement type of the wind farm operation and maintenance calculation task; Generate the first-step length adjustment factor and the second-step length adjustment factor according to the task priority and data requirement type. Based on the first step length adjustment factor and the second step length adjustment factor, the preset jump step length obtained by rounding up is optimized, and the optimized preset jump step length is output.

7. The wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in claim 1, characterized in that, The method for evaluating the set of candidate rollback nodes to obtain the rollback cost index for each candidate rollback node includes: Obtain the set of possible paths for each candidate fallback node; The optional path quantification index is obtained by weighted calculation of the path length and node load change rate of the optional path set. The optional path hop count index is the average hop count index of all paths in the optional path set. The optional path failure probability index is obtained by weighted fusion calculation of the probability accumulation of the historical failure probability of nodes and the historical failure probability of links of all paths in the optional path set.

8. The wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in claim 7, characterized in that, The method for reconstructing the first jump path using the first fallback node as the successor node includes: Obtain a set of possible paths with the first fallback node as the successor node, where the first fallback node is the candidate fallback node with the lowest fallback cost index in the candidate fallback node set. The first jump path is obtained by sorting the set of optional paths according to the fallback cost index.

9. A wind farm hierarchical operation and maintenance management system based on multi-level monitoring, characterized in that, The steps for implementing the wind farm hierarchical operation and maintenance management method based on multi-level monitoring as described in any one of claims 1 to 8, wherein the wind farm hierarchical operation and maintenance management system based on multi-level monitoring comprises: The monitoring transmission network establishment module is used to establish a multi-level monitoring transmission network based on a set of monitoring terminal nodes; The initial path transmission module is used to transmit the initial path from the multi-level monitoring transmission network according to the wind farm operation and maintenance calculation task, and triggers a fallback decision mechanism when a fault is detected in the next path node. The rollback decision module is used by the rollback decision mechanism to construct a set of target reachable nodes with reference to the target node of the wind farm operation and maintenance calculation task. The node filtering module is used to analyze and filter the target reachable node set to obtain a candidate fallback node set. The rollback cost evaluation module is used to evaluate the set of candidate rollback nodes to obtain the rollback cost index of each candidate rollback node, including the optional path quantification index, the optional path jump count index, and the optional path failure probability index. The data retransmission module is used to identify the first rollback node using the rollback cost index, reconstruct the first jump path with the first rollback node as the successor node, and re-execute the operation and maintenance data transmission of the wind farm operation and maintenance calculation task according to the first jump path.

Images