Multi-target collaborative intelligent management and control method and system for electric power materials in multi-scheduling scene

Abstract

The invention belongs to the technical field of material full-process management and control, and provides an electric power material multi-target cooperative intelligent management and control method and system in a multi-scheduling scene, and the method comprises the steps: carrying out the multi-stage omnidirectional checking of electric power materials: carrying out the adaptive sorting of the electric power materials through considering the signal intensity, phase stability, polarization matching degree and misreading compensation factors, hierarchical scanning is realized; a storage location accurate distribution step: determining a storage location distribution scheme according to various scheduling scene requirements; an electric power material digital twinborn coupling verification step: identifying the electric power material, determining a space coordinate of the corresponding electric power material in the digital twinborn model, performing coupling verification in combination with an identification result, the space coordinate and a storage location distribution scheme, and updating storage management and control information; and a multi-target collaborative optimization scheduling step: comprehensively considering the multi-source data of the electric power materials to carry out risk assessment, and carrying out multi-target collaborative optimization. The efficiency of the whole electric power material inspection, storage and distribution process is improved.

Description

Technical Field

[0001] This invention belongs to the field of whole-process material management and control technology, specifically relating to a multi-objective collaborative intelligent management and control method and system for power materials under multiple scheduling scenarios. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Power equipment is a crucial asset for ensuring the stable operation of the power grid, making its management and control extremely important. However, the management and control of power equipment throughout its entire lifecycle involves numerous stages, such as procurement, warehousing, storage, inventory, warehouse management, dispatching, and operation and maintenance. The sheer number of stages, and the fact that some stages still require human intervention, leads to problems in the management and control of power equipment, including low efficiency, low accuracy, and discrepancies between the reserves and demand of certain equipment in some scenarios.

[0004] Specifically, there are also certain problems in each step: In the warehousing process, the existing power material storage location allocation scheme does not distinguish the priority difference between maintenance operations and emergency repair operations. Emergency repair materials often suffer from problems such as remote storage locations or being blocked by other materials due to unreasonable storage location allocation, resulting in retrieval delays and failing to meet the needs of rapid power failure handling.

[0005] In the inventory process of power materials, RFID inspection robots based on fixed readers are susceptible to signal interference from metal materials in the warehouse, which contains a large number of static materials stored in multiple dimensions. This results in a shortened identification distance, reduced accuracy, and blind spots, making it easy to miss readings. The tags of materials in the upper and deeper parts of the shelves are difficult to read effectively, leading to a serious disconnect between inventory data and actual location. On the other hand, when identifying power equipment in batches during the inbound and outbound process, the concentrated identification of multiple tags is prone to signal interference and missed readings. Furthermore, the different types of equipment, urgency of inbound and outbound situations, and locations in different power warehouse areas increase the difficulty of processing identification signals.

[0006] In the storage and management of power materials, there are many types of power materials, and the safety risks are characterized by differentiation, suddenness, and transmissibility. The existing management model has problems such as lagging regular inspections, single risk assessment, and fragmented traceability data. Moreover, some materials have special storage constraints, such as high-voltage equipment requiring independent insulated storage and precision instruments requiring constant temperature and moisture-proof environment. Existing solutions often fail to balance the priorities of multiple factors and storage constraints when allocating storage space, which can easily lead to contradictions such as meeting storage requirements but low operational efficiency or pursuing efficiency while neglecting storage safety.

[0007] In the power material dispatching process, we are often in a passive mode, and there are also multi-scenario intelligent dispatching across regions (i.e., across different storage areas). Existing solutions generally pursue the shortest time or the shortest path, which can easily lead to contradictions such as short path congestion and time consumption, high speed and high energy consumption, and high safety and low efficiency. Summary of the Invention

[0008] To address the aforementioned problems, this invention proposes a multi-objective collaborative intelligent management and control method and system for power materials under multiple scheduling scenarios. This invention improves key aspects such as warehousing, inventory, storage management, and scheduling, thereby enhancing the accuracy and efficiency of power material management and control.

[0009] According to some embodiments, the present invention adopts the following technical solution: A method for multi-objective collaborative intelligent management and control of power resources in multi-scheduling scenarios includes the following steps: The steps for precise allocation of storage locations are as follows: Based on the needs of various scheduling scenarios, determine the types and quantities of required power materials, generate demand instructions, and determine the storage priority of power materials under storage constraints based on the warehousing impact coefficient of the power materials to be allocated and the scheduling scenario type, thereby determining the warehousing location allocation strategy. Multi-level omnidirectional inventory steps for power materials: Based on the weighted sum of the signal strength, phase stability, polarization matching degree, and missed reading compensation factor of the power materials tags, determine the adaptive sorting factor, and perform omnidirectional hierarchical inventory based on the priority of the adaptive sorting factor. The steps for coupling verification of digital twin of power materials are as follows: Based on the image data of the acquired storage area, the semantic information of power materials is identified, the spatial coordinates of the corresponding power materials in the storage digital twin model are determined, and the corresponding storage location is associated with it. The coupling verification is carried out by combining the identified semantic information, spatial coordinates and storage location allocation strategy, and the storage management information is updated. Multi-objective collaborative optimization scheduling steps: Taking into account the physical state, environmental impact, historical operating load and life cycle data of power materials, risk assessment is carried out to predict the scheduling scenarios of power materials. Combined with the storage management information of each storage area, multi-objective collaborative optimization is carried out with the goal of simultaneously achieving the shortest path, the shortest total scheduling time and the lowest scheduling energy consumption, and scheduling strategies are pre-generated.

[0010] A multi-objective collaborative intelligent management and control system for power materials under multiple scheduling scenarios includes: The warehouse location precision allocation unit is used to determine the type and quantity of required power materials according to the needs of various scheduling scenarios, generate demand instructions, and determine the storage priority of power materials under storage constraints based on the warehousing impact coefficient of the power materials to be allocated and the scheduling scenario type, thereby determining the warehousing location allocation strategy. The multi-level omnidirectional inventory unit for power materials is used to determine the adaptive sorting factor based on the weighted sum of the signal strength, phase stability, polarization matching degree and missed reading compensation factor of the power materials tags in the warehouse, and to perform omnidirectional hierarchical inventory based on the priority of the adaptive sorting factor. The digital twin coupling verification unit for power materials is used to identify the semantic information of power materials based on the image data of the acquired storage area, determine the spatial coordinates of the corresponding power materials in the storage digital twin model, associate them with the corresponding storage location, and perform coupling verification by combining the identified semantic information, spatial coordinates and storage location allocation strategy to update the storage management information. The multi-objective collaborative optimization scheduling unit is used to comprehensively consider the physical state, environmental impact, historical operating load and life cycle data of power materials, conduct risk assessment, predict the scheduling scenarios of power materials, and combine the storage management information of each storage area to achieve multi-objective collaborative optimization with the goal of achieving the shortest path, the shortest total scheduling time and the lowest scheduling energy consumption, and pre-generate scheduling strategies.

[0011] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the steps in the above method.

[0012] An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the steps in the method described above.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention innovatively proposes a multi-objective collaborative intelligent management and control method for power materials under multiple scheduling scenarios. It coordinates and optimizes multiple links such as inventory, storage location allocation, warehouse management and scheduling, effectively improving the efficiency and accuracy of power material management. Moreover, it can guide the procurement, storage, scheduling and distribution of power materials in an orderly manner according to different scheduling scenarios, which is of great significance for the intelligent management of the entire chain of power material storage and distribution.

[0014] This invention innovatively proposes a method for precise allocation of storage locations. During location allocation, it reserves reasonable storage space for support materials or prioritizes the allocation of locations with convenient scheduling, thereby enhancing the overall guarantee capability of regional power materials and helping to build an efficient and interconnected power material storage network. By differentiating between maintenance and emergency repair operations, it dynamically adjusts the weight of each inbound impact coefficient. Under the premise of meeting storage constraints, it selects storage locations based on storage priority, balancing storage constraints and priorities, and accurately adapting to the core needs of different operational scenarios. This achieves dual protection of storage safety and operational efficiency, improving the accuracy, efficiency, and emergency response capability of power material storage management.

[0015] This invention innovatively proposes a multi-level omnidirectional inventory method for power materials. Based on the weighted sum of signal strength, phase stability, polarization matching degree, and missed reading compensation factor of the tags on the power materials entering the warehouse, an adaptive sorting factor is determined. Hierarchical scanning is then performed according to the priority of the adaptive sorting factor. Furthermore, it abandons the directional scanning method and designs a completely new omnidirectional inventory method. Omnidirectional inventory is achieved through movement and rotation. During rotation, the forward and reverse limit rotation time is matched using a speed-time-current correspondence table. A friction compensation coefficient defined by dust concentration and limit rotation angle, along with the cumulative motor running time, is introduced to quantify the rotation time compensation. The final rotation time is dynamically corrected through this time compensation, avoiding initial intermediate position positioning deviations caused by sensor failure.

[0016] This invention innovatively proposes an omnidirectional inventory supplementation scanning method for power materials. By self-optimizing the positioning of easily missed areas, it solves the drawbacks of blind supplementation scanning and improves the targeting of supplementation scanning. It dynamically determines the offset threshold based on the type of power material and the initial missed scanning angle. Combining the offset threshold with the type of power material, it adaptively adjusts the scanning speed to solve the problems of inefficiency caused by over-scanning and inventory errors caused by under-scanning. It achieves the dual goals of coverage without blind spots and scanning without redundancy. It realizes the orientation of the supplementation range, the dynamic adjustment of the threshold, and the adaptive scanning speed, thereby improving the coverage integrity of missed areas, the accuracy of inventory data, the supplementation scanning efficiency, and the system adaptability flexibility.

[0017] This invention innovatively proposes a digital twin coupling verification method for power materials. Based on image data of the storage area, power materials are identified, a digital twin model is constructed, the actual spatial coordinates of the corresponding power materials in the three-dimensional space of the storage area are determined, and the model is associated with the corresponding storage location. The identification results, actual spatial coordinates, and storage location allocation scheme are combined for coupling verification to update the storage management information. This method can achieve accurate matching and verification of inventory and storage information on a regular or dynamic basis, ensuring the accuracy of power material information and better serving subsequent scheduling.

[0018] This invention innovatively proposes a multi-objective collaborative optimization scheduling method for power materials. It comprehensively considers the characteristics of the physical state, environmental impact, operating load, and life cycle data of power materials, conducts risk assessment, predicts scheduling scenarios for power materials, and combines storage management information from various storage areas. The method achieves multi-objective collaborative optimization with the goal of simultaneously achieving the shortest path, the shortest total scheduling time, and the lowest scheduling energy consumption. It generates scheduling strategies that are adapted to the safety characteristics of different materials such as transformer oil leakage and cable aging, realizing diversified management and control, identifying hidden risks in advance, effectively reducing the consumption of management and control resources, and improving scheduling efficiency and response speed.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0020] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0021] Figure 1 This is a flowchart of a multi-objective collaborative intelligent management and control method for power materials in a multi-scheduling scenario, as one embodiment. Figure 2 This is a flowchart illustrating the precise allocation steps for warehouse storage locations in one embodiment. Figure 3 This is a flowchart illustrating the steps of a multi-level omnidirectional inventory of power materials in one embodiment. Figure 4 This is a flowchart of the digital twin coupling verification steps for power materials according to one embodiment; Figure 5 This is a flowchart of the multi-objective collaborative optimization scheduling steps in one embodiment. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0023] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0025] Where there is no conflict, the embodiments and features described in this application may be combined with each other.

[0026] Example 1 A method for multi-objective collaborative intelligent management and control of power materials in multi-scheduling scenarios, such as Figure 1 As shown, it includes the following steps: Step S1, Precise Allocation of Warehouse Locations: Based on the needs of various scheduling scenarios, determine the types and quantities of required power materials, generate demand instructions, and determine the storage priority of power materials under storage constraints based on the warehousing impact coefficient of the power materials to be allocated and the scheduling scenario type, thereby determining the warehousing location allocation strategy. Step S2, Multi-level Omnidirectional Inventory of Power Materials: Based on the weighted sum of the signal strength, phase stability, polarization matching degree, and missed reading compensation factor of the power materials tags, determine the adaptive sorting factor, and perform omnidirectional hierarchical inventory based on the priority of the adaptive sorting factor. Step S3, Digital Twin Coupling Verification Step for Power Materials: Based on the acquired image data of the storage area, identify the semantic information of the power materials, determine the spatial coordinates of the corresponding power materials in the storage digital twin model, associate them with the corresponding storage location, and perform coupling verification by combining the identified semantic information, spatial coordinates and storage location allocation strategy to update the storage control information. Step S4, Multi-objective Collaborative Optimization Scheduling Step: Taking into account the physical state, environmental impact, historical operating load, and life cycle data of power materials, conduct risk assessment, predict the scheduling scenarios of power materials, and combine the storage management information of each storage area to achieve multi-objective collaborative optimization with the goal of simultaneously achieving the shortest path, the shortest total scheduling time, and the lowest scheduling energy consumption, and pre-generate scheduling strategies.

[0027] In this embodiment, as Figure 3 As shown, in this embodiment, step S1 includes: Step S101: Based on the maintenance and emergency repair needs, determine the required power supplies and their current inventory. If the current inventory is lower than the set threshold, issue a distress call. Step S102: Obtain the warehousing impact coefficient of the power materials to be allocated, including weight, volume, frequency of entry and exit, storage requirements, and distress requests from other warehouses; Step S103: Determine the weight of each inbound impact coefficient according to the operation type to obtain the storage priority of power materials; Step S104: Under the constraint of meeting storage requirements, select storage locations according to the storage priority of power materials to obtain a storage location allocation scheme.

[0028] During the spring / autumn maintenance period, if a maintenance request is received, the inventory of electrical materials in the current warehouse will be checked and sorted out using inventory equipment to determine the required electrical materials and their current inventory quantity. The results can then be fed back to the integrated processing center of this warehouse and other warehouses.

[0029] Then, based on the maintenance needs and the characteristics of the power materials, and according to the established storage location allocation plan, the handling equipment is mobilized to move the power materials to the optimal storage location.

[0030] If the current inventory level is lower than the set threshold, it indicates that the warehouse is short of the required power supplies. The integrated processing center will then issue a distress call based on the inventory status of other warehouses and may also provide purchase information alerts.

[0031] If no maintenance work request for spring / autumn inspection is received, and it is currently in the spring inspection (March, April, and May each year) or autumn inspection (September, October, and November each year) period, the handling procedure is the same as above and will not be repeated.

[0032] However, if it is not during the maintenance period and no maintenance work request for spring / autumn maintenance has been received, then it is only necessary to take inventory and send the inventory results to other warehouses.

[0033] When an emergency repair request is received, it should be given priority regardless of whether it is during the spring / autumn maintenance period.

[0034] Specifically: Upon receiving an emergency repair request, immediately trigger the power material inventory process to determine the required power materials and their current inventory in the warehouse, in order to quickly verify whether the warehouse has the power materials required for the repair.

[0035] If the current inventory is not lower than the set threshold, it means that the warehouse has sufficient power supplies. The integrated processing center will provide the specific storage location information of the required power supplies in a timely manner. Based on the established storage location allocation plan, and combined with the existing number of personnel and their distribution locations in the warehouse, the center can also mobilize handling equipment to move the power supplies to the best storage location.

[0036] If the current inventory level is lower than the set threshold, it indicates that there is a shortage of power supplies in this warehouse. The integrated processing center will combine the inventory data and distance information of other warehouses for comprehensive analysis and send a distress request to the nearest warehouse.

[0037] In this embodiment, the warehouse location allocation scheme takes into account five factors that affect the entry of power materials into the warehouse: weight, space (including regular and irregular power materials), frequency of entry and exit (considering spring / autumn inspections and emergency repairs), storage conditions, and the distress needs of other warehouses.

[0038] Once the power supplies to be allocated are identified (which may be power supplies in the current warehouse that need to be relocated, or newly arrived power supplies that need to be put into storage), weight and volume data are collected using weight sensors and image sensors. Then, combined with the frequency of power supplies entering and leaving the warehouse, storage requirements, and distress requests from other warehouses, storage priority is evaluated according to dynamically adjusted weight rules to guide the allocation of storage locations.

[0039] First, the impact coefficient of warehousing is quantified into specific assessable levels or values, specifically: (1) Weight (W): Generally, the greater the weight of a single item, the more difficult it is to handle.

[0040] Therefore, weight class thresholds are set, namely a first weight threshold and a second weight threshold, with the first weight threshold being less than the second weight threshold; If the weight of the power materials to be allocated is less than or equal to the first weight threshold, it is considered a light power material and quantified as 1. If the weight of the power materials to be allocated is less than or equal to the second weight threshold and greater than the first weight threshold, it is considered as medium power materials and quantified as 2. If the weight of the power materials to be allocated is greater than the second weight threshold, it is considered a heavy power material and quantified as 3.

[0041] (2) Space (V): Generally, the more space occupied, the higher the requirements for shelving.

[0042] Power resources are categorized into regular power resources and irregular power resources. The difference between regular and irregular power resources lies in their spatial quantification methods. For standard electrical equipment, the length, width, and height are directly measured to calculate the volume, and then it is quantified into three levels: small, medium, and large, based on its actual size. For irregularly shaped electrical equipment, the space occupation is quantified based on standard storage units (SSUs). The required number of SSUs is assessed through three-dimensional data or manual experience, while also considering whether the electrical equipment can be stacked and whether special operations are required (such as hoisting or reserving a safety area). Appropriate penalty items are added to adjust the space quantification results, and finally, it is also divided into three levels: small, medium, and large.

[0043] Specifically: (2-1) The volume V of regular electrical materials is calculated directly from their length, width, and height; At the same time, volume level thresholds are set, namely a first volume threshold and a second volume threshold, and the first volume threshold is less than the second volume threshold; If the volume of the power supply to be allocated is less than or equal to the first volume threshold, it is considered a small power supply and quantified as 1. If the volume of the power materials to be allocated is less than or equal to the second volume threshold and greater than the first volume threshold, it is considered as medium power materials and quantified as 2. If the volume of the power supply to be allocated is greater than the second volume threshold, it is considered a large power supply and quantified as 3.

[0044] (2-2) For irregularly shaped power materials (such as cable reels, insulator strings, irregularly shaped fittings, transformers, etc.), the volume cannot be simply calculated using length, width and height, but their space occupation has a huge impact on storage efficiency.

[0045] Therefore, the most basic storage space in a warehouse (such as a pallet location, a storage compartment, or a floor area) is defined as a standard storage unit; for example, 1 standard storage unit = 1 pallet location of 1m × 1.2m, or a storage compartment of 1m³.

[0046] Then, by acquiring the 3D point cloud data of the power supplies to be allocated, the number of standard storage units they occupy is determined, resulting in the following volume: V = SSU × (1 + SP + HP); Wherein, SSU is the number of standard storage units occupied, SP is the stacking penalty, SP=0 when the power materials to be allocated can be safely stacked, and SP=0.2 when they cannot be stacked, which means that the vertical space above is wasted; HP is the operation penalty, HP=0 when the power materials to be allocated do not require special operation, and HP=0.1~0.3 when special operation is required (such as when a transformer needs to be hoisted and a safe zone needs to be left around it).

[0047] Finally, volume quantization can be performed based on the set volume level threshold.

[0048] Alternatively, it can be divided according to the number of SSUs occupied, such as setting a threshold for the number of occupied levels, namely a first threshold for the number of occupied and a second threshold for the number of occupied, and the first threshold for the number of occupied (e.g., 1) is less than the second threshold for the number of occupied (e.g., 3). If the number of occupied power materials (SSU) is less than or equal to the first occupied quantity threshold, it is considered a small power material and quantified as 1. If the number of occupied power materials to be allocated (SSU) is less than or equal to the second occupied quantity threshold and greater than the first occupied quantity threshold, it is considered as medium power materials and quantified as 2. If the number of occupied power materials (SSU) is greater than the second occupation threshold, it is considered a large power material and quantified as 3.

[0049] (3) Inbound / outbound frequency (F): The higher the inbound / outbound frequency, the higher the requirements for operational efficiency. During the spring and autumn inspections, the inbound / outbound frequency of electrical materials such as insulating gloves, filters, transformers, capacitors, surge arresters, and cables will increase significantly.

[0050] F = Total number of inbound and outbound transactions within a historically defined time period (e.g., 30 days / 90 days).

[0051] To eliminate differences in magnitude, normalization can be performed: F n =(FF min ) / (F max -F min ), F min and F max These are the minimum and maximum values.

[0052] Alternatively, levels can be defined, and frequency level thresholds can be set, namely a first frequency threshold and a second frequency threshold, wherein the first frequency threshold is less than the second frequency threshold. If the frequency of entry and exit of the power materials to be allocated is less than or equal to the first frequency threshold, they are considered low power materials and quantified as 1. If the inbound and outbound frequency of the power materials to be allocated is less than or equal to the second frequency threshold and greater than the first frequency threshold, it is regarded as medium-power materials and quantified as 2. If the inbound and outbound frequency of the power materials to be allocated is greater than the second frequency threshold, it is regarded as high-power materials and quantified as 3.

[0053] (4)Storage condition (C): The requirements of the storage condition directly affect the choice of storage location. This is a reverse index. The more demanding the condition, the fewer available storage locations, and usually special areas are required for storage. It is quantified as: ordinary (1), moisture-proof (2), constant temperature and humidity (3), special dangerous goods (4). The higher the value, the more restrictions there are.

[0054] (5)The rescue needs of other warehouses (E): The warehouse has realized cross-regional transfer and multi-warehouse collaboration. Therefore, the choice of storage location for power materials needs to consider the rescue needs of other warehouses. In this case, there are some power materials with low inbound and outbound frequencies usually, but become urgent due to the rescue needs of other warehouses. In this case, even if the inbound and outbound frequency F is not high, it needs to be placed in an easily accessible location.

[0055] It is quantified as: regular inventory (0), reserved by rescue needs (1), rescue needs require execution within 24 hours (2); the rescue site is in place and waiting for power materials (3).

[0056] Next, perform dynamic adjustment of the weight coefficients. In this embodiment, determine the operation types, including maintenance operations, mixed operations, and emergency repair operations; among them, spring inspection / autumn inspection belongs to the maintenance operation mode. Define the operation coefficient k according to the operation type. The operation coefficient k is a continuous variable between 0 and 1: When k = 1: It represents the emergency repair operation mode.

[0057] When k = 0: It represents the regular maintenance operation mode.

[0058] When 0 < k < 1: It represents the mixed mode. The higher the k value, the stronger the urgency of the emergency repair.

[0059] k = (the reciprocal of the required response time - the reciprocal of the response time of the maintenance operation mode) / (the reciprocal of the response time of the emergency repair operation mode - the reciprocal of the response time of the maintenance operation mode); where the time is in minutes.

[0060] Then adjust the weight ratio of each inbound influence coefficient.

[0061] Generally, the frequency of inbound and outbound operations has the highest weight, as it is the primary factor affecting operational efficiency; storage conditions, as a hard constraint, have the next highest weight; weight and space have relatively low weights, mainly affecting handling and space utilization; the weight of other warehouse distress calls increases as the urgency of the repair increases.

[0062] (a) Inbound / outbound frequency weight γ: γ(k)=γ0+k*(γ1-γ0)=0.20+k*(0.40-0.20)=0.20+0.20k; Wherein, γ0 is the basic weight of the frequency of entry and exit from the warehouse under the regular maintenance operation mode; γ1 is the basic weight of the frequency of entry and exit from the warehouse under the emergency repair operation mode.

[0063] (b) Storage condition weight δ: δ(k)=δ0+k*(δ1-δ0)=0.45+k*(0.25-0.45)=0.45-0.20k; Wherein, δ0 is the basic weight of storage conditions for the regular maintenance operation mode; δ1 is the basic weight of storage conditions for the emergency repair operation mode.

[0064] (c) Weight weight α: α(k) = 0.15, which remains unchanged.

[0065] (d) Spatial weight β: β(k) = 0.1-0.05k.

[0066] (e) The weight of the distress call demand from other warehouses is ε(k) = 0.1 + 0.05k.

[0067] (f) Finally, calculate the storage priority L of power materials: L=α(k)*W+β(k)*V+γ(k)*F+ε(k)*E-δ(k)*C; L=0.15*W+(0.1-0.05k)*V+(0.20+0.20k)*F+(0.1+0.05k)*E-(0.45-0.20k)*C; α+β+γ+δ+ε=1.

[0068] Here, α*W+β*V indicates that the greater the weight and space occupied, the lower and more robust the shelf should be placed, or the closer to the exit to shorten the handling distance, thus making a positive contribution to L.

[0069] γ*F indicates that the higher the frequency of inbound and outbound operations, the greater the positive contribution to L, which is the key to improving efficiency.

[0070] ε(k)*E represents the power supplies that receive distress requests from other warehouses and are moved closer to the exit to shorten the handling distance, contributing positively to L.

[0071] -δ*C indicates that the more stringent the storage requirements (the larger the C value), the greater the negative contribution to L. This is because storage requirements restrict electrical resources from being placed in ordinary, convenient areas, and they must instead enter specific areas, which may not be on the optimal route.

[0072] In this embodiment, the storage condition is a hard constraint. Therefore, under the constraint of meeting the storage requirements, the storage areas that meet the storage requirements are selected, and then the storage locations are sorted and selected according to the storage priority of power materials, so as to obtain the storage location allocation scheme.

[0073] The storage location allocation rules are as follows: (1) High-end power materials: prime locations (doorway, ground floor, picking area).

[0074] (2) Zhong L power materials: ordinary storage location.

[0075] (3) Low-L power materials: high-rise shelves, deep storage areas.

[0076] Any application with special storage requirements should have those requirements met first, and then sorted by L within its dedicated area.

[0077] As a typical example, the scenario is as follows: after a rainstorm or typhoon, there are multiple faults in the power grid, and power supplies are urgently needed for repair.

[0078] The scenario is: emergency repair work, K=1.

[0079] L=0.15*W+0.05*V+0.40*F+0.15E-0.25*C.

[0080] Among them, the surge arrester: L=0.15*2+0.05*1+0.4*1+0.15*0-0.25*2=0.3; Transformer: A large transformer that cannot be stacked, requires hoisting, and is required to be installed within 24 hours of receiving a distress call. The SSU is 4, then L = 0.15*2 + 0.05*3 + 0.4*1 + 0.15*2 - 0.25*3 = 0.4; Therefore, the transformer has a high L value and should be placed in the prime storage location.

[0081] like Figure 3 As shown, step S2 specifically includes: S201. Obtain multiple tag data within the warehouse area; S202. By weighted summation of signal strength, phase stability, polarization matching degree and missed reading compensation factor, an adaptive sorting factor for multiple tag data is obtained. The weights of signal strength, phase stability, polarization matching degree and missed reading compensation factor are adjusted according to the different types of power equipment in the power warehouse. S203. Sort the data according to the adaptive sorting factor, and scan the label data with the adaptive sorting factor value higher than the preset value first, and then scan the other label data.

[0082] In this embodiment, in step S201, an infrared sensor can be used to automatically trigger a scanning task, and an RFID reader can be used to quickly identify the physical ID electronic tags of power materials, conduct real-time inventory of power materials entering and leaving the warehouse, and upload key information.

[0083] Of course, in other embodiments, the tag data can be obtained from other data, such as robot inventory equipment.

[0084] In some embodiments, during tag recognition, adaptive adjustment of tag response timing and signal characteristics enables orderly response and efficient differentiation of multiple tags within the same recognition area, prioritizing the scanning of high-priority tags. Combined with a signal strength discrimination and tag response status dynamic management mechanism, the probability of signal collisions between multiple tags can be effectively reduced, improving the completeness and accuracy of batch recognition. Dynamic adjustment of the reader's transmission power suppresses long-distance tag interference. (Number of tags) The estimate is: ; in, Number of idle time slots; This represents the total number of time slots. This represents the number of successful time slots.

[0085] Optimal frame length Adjusted to: ; in, Find the minimum value of the expression; This refers to any frame length value from the candidate set of selectable frame lengths for the system, such as 64, 128, 256, 512, etc. The initial frame length is 64 time slots; the maximum frame length is 512 time slots; and the tag recognition rate is ≥99%.

[0086] In this embodiment, step S202 distinguishes tags within the same identification area by quantifying and scoring tag feature parameters. The selected feature factors are signal strength RSSI, carrier phase stability (Phase), polarization matching degree (pol), and historical unread tag markers (misstag). Specifically, RSSI is used to differentiate between near and far obstructions, improving identification priority; Phase stability (Phase) is used to determine tag response reliability and location visibility; polarization matching degree (pol) is used to determine the spatial attitude adaptation between the antenna and the tag; and the historical unread tag markers (misstag) are used to prioritize compensation for tags not identified in the previous round.

[0087] In this embodiment, the multi-label adaptive ranking factor in step S203 : ; in, This is the standard value for signal strength. This is the standard value for phase stability; This represents the standard value for polarization matching. This is the standard value for the missed reading compensation factor; , , and These are the weights corresponding to signal strength, phase stability, polarization matching degree, and missed read compensation factor, respectively.

[0088] When determining the standard value, the collected feature factors are normalized: The method for normalizing RSSI signal strength to a standard signal strength value S (0~1) involves linear scaling and truncation. A minimum and maximum readable signal value (RSSI_min and RSSI_max) are set, and the signal is proportionally converted to a range of 0~1: the closer to the maximum value, the closer to 1. Optional. → ,For example dBm, dBm.

[0089] Phase stability is normalized to a standard phase stability value Ph (0~1) using a phase fluctuation mapping method. The stability or rate of change of the tag signal's phase is calculated, and then compressed to 0~1 using a function. The more stable the phase (the smaller the error), the higher the score. Optional, → (Map phase stability or phase change rate to [0, 1], the higher the stability, the closer it is to 1).

[0090] The method for normalizing polarization matching to a standard value Po (0~1) involves using directional consistency mapping to determine the degree of matching between the tag orientation and the antenna polarization orientation. Mapping the matching degree to 0~1: the more consistent the orientation, the higher the readout efficiency and the higher the score. Optional, → (Polarization matching degree is mapped to [0, 1]).

[0091] The missed read compensation factor is normalized to a standard value M (0~1) using the following method: weighted by the number of missed reads, the most recent unread count (miss_count) of a tag is counted, and the priority of unidentified tags is increased: the more missed reads, the closer the value is to 1. Optional, → (The weight of unread history is closer to 1 the more times it has not been identified recently;) (regularity coefficient).

[0092] Standardized weight normalization: ,and .

[0093] Of course, the weighting coefficients can be set according to different specific scenarios. This is just one example, illustrating the inbound and outbound operations of a typical power warehouse (where most materials are non-metallic packaged and moderately concealed). Optional... Emergency / replacement priority (urgent delivery or critical spare parts, priority should be given to identifying those that have not been read for a long time or those with critical labels); optional. In this scenario, the MissTag weight is increased to prevent critical materials from being missed for a long time.

[0094] Sort all tags by adaptive sorting factor Sort, take the first (or (This is a fast-track read, with subsequent reads in batches.) A standard value for the missed read compensation factor is specified. High value (e.g.) If a tag fails to read successfully on the first attempt, an additional retry round and an adjustment of the transmission power will be arranged.

[0095] In this embodiment, when a large number of strong signal clutter (a large number of unmatched materials with extremely high RSSI) is detected at a distance, the transmission power is reduced and a near-field priority reading strategy is activated to suppress far-field interference.

[0096] In addition, the omnidirectional inventory referred to in this embodiment refers to the use of an omnidirectional inventory device to achieve omnidirectional inventory, which is suitable for omnidirectional barcode scanning inventory scenarios in power material warehouses, especially for DC motors without Hall effect sensors / encoders, and involves two parts: dynamic speed-assisted positioning and adaptive protection of the motor in the omnidirectional inventory device.

[0097] In this embodiment, it specifically includes: Based on the rotational current of the motor at the set speed, the forward and reverse limit rotation time is obtained by matching the constructed speed-time-current correspondence table. The friction compensation coefficient is defined based on the dust concentration and the forward and reverse limit rotation angles. The time compensation amount is determined by combining the cumulative running time of the motor. The final rotation time is obtained based on the forward and reverse limit rotation time and the time compensation amount. The initial intermediate position is defined as the position after the motor has rotated T / 2 in the opposite direction from any physical limit point, and T is the final rotation time. The motor is controlled to start rotating from the initial intermediate position, and the rotation current is obtained. When the motor reverses direction, the protection threshold is updated based on the commutation peak current, and the current rotational current is compared with the updated protection threshold. If the motor does not reverse direction, the current rotational current is compared with the original protection threshold to determine whether a fault has occurred. If a fault occurs, the fault type is determined based on the time point at which the current rotational time is located.

[0098] Because the materials in the power supply warehouse are placed in a compact manner, the radio frequency antenna needs to be rotated in both clockwise and counterclockwise directions from the initial center position to cover the inventory area. Therefore, the initial center position is determined first, which includes the following:

[0099] Step S21: Initial calibration stage, three-speed calibration to adapt to warehouse voltage fluctuations.

[0100] Specifically, it includes: (1) When using it for the first time, manually trigger the motor to rotate to the clockwise physical limit and record the initial trigger signal.

[0101] (2) Control the motor to rotate from the clockwise physical limit to the counterclockwise physical limit at different speeds, and obtain the clockwise and counterclockwise limit rotation time and rotation current.

[0102] The motor is controlled to run at a set low speed (5° / s), rotating from the clockwise physical limit to the counterclockwise physical limit, and a timer records the rotation time for both directions. T 1. Sample the motor current during rotation. I 1; Control the motor to run at a set medium speed (30° / s), repeat the above operation, and record the clockwise and counterclockwise limit rotation time. T 2 and motor current I 2; Control the motor to run at a set high speed (60° / s), and record the clockwise and counterclockwise limit rotation times. T 3 and motor current I 3; This generates a table corresponding to rotational speed, time, and current.

[0103] (3) Considering that after prolonged use of the motor, the friction of the drive shaft increases, affecting the rotational speed, the Stribeck friction model is used. and motor speed Derive the relationship between the rotational speed n and the rotational speed n; where, It is the total frictional torque. For sliding friction torque, The viscous damping coefficient is... Angular velocity, It is the output torque of the motor. It is the load torque.

[0104] The derivation process is as follows: ; ; Where n is the motor speed, It is the torque constant. It is the back electromotive force constant; This refers to the motor's output power. This is the output current of the motor.

[0105] Therefore, it can be concluded that with the use of motors, Increasing the output torque of the same motor will result in The decrease in rotational speed n necessitates the definition of a friction compensation coefficient. K f It is used to correct the situation where the torque remains the same but the speed decreases due to increased friction in the drive shaft.

[0106] ; Where t is the cumulative running time; The rotation angle is for clockwise and counterclockwise limits. This is the rotation speed under clean conditions. This is the actual speed difference. C represents the dust concentration. =0.05 is a proportionality constant.

[0107] Finally, the compensation amount for the clockwise and counterclockwise limit rotation time caused by dust friction per unit operating time is defined as: This corrects the clockwise physical limit rotation time.

[0108] Step S22, Daily Positioning Stage: Utilizing the positive correlation between DC motor current and speed, the speed is indirectly matched by sampling the real-time current, without the need for additional voltage detection. Specifically, it includes: (1) Before each inventory count, control the motor to run clockwise and counterclockwise for 2 seconds at the set speed, and collect the motor current during the rotation.

[0109] (2) Using a speed-time-current correspondence table, the forward and reverse limit rotation time corresponding to the current speed is determined by matching the motor current.

[0110] (3) Based on the defined friction compensation coefficient and the cumulative running time of the motor, through The time compensation amount is calculated, and the final rotation time T is obtained by adding the time compensation amount to the forward and reverse limit rotation time.

[0111] This is because the time in the corresponding table is the rotation time of the inventory device when it leaves the factory, which is the initial time. As the motor runs for a longer time, and as friction increases, the rotation angle will decrease for the same motor current and rotation time. Therefore, the time compensation is added to obtain the final rotation time.

[0112] (4) Control the motor to rotate in the opposite direction from the current physical limit position. The rotation time is T / 2. After stopping, it will be the initial intermediate position.

[0113] Step S23, Periodic calibration stage, automatic limit calibration to adapt to mechanical wear.

[0114] Specifically, it includes: (1) After each set number of inventory cycles (e.g., 10 times) is completed, a calibration command is automatically triggered. The motor rotates from its current position to the clockwise physical limit and then to the counterclockwise physical limit. The clockwise and counterclockwise limit rotation time T is recorded again. 新 .

[0115] (2) Compare T 新 If the deviation from the historical time T exceeds the set threshold, the speed-time-current correspondence table is updated to cover the impact of limit position caused by mechanical wear.

[0116] In this embodiment, after determining the initial intermediate position, the motor is controlled to start rotating from the initial intermediate position. During the rotation process, the motor state is judged and adaptive protection is performed.

[0117] Specifically, it includes the following:

[0118] (1) Set protection threshold.

[0119] When the motor is not commutated, the protection threshold is a multiple of the rated current; When the motor reverses direction, adjust the protection threshold to a multiple of the commutation peak current.

[0120] Specifically: First, the commutation spike current is determined based on Kirchhoff's voltage law; Kirchhoff's voltage law is: ; in, Rated voltage, =12V; For armature inductance, , R a For armature resistance, R a =1.265Ω; The rate of change of current, The change in current For time intervals; For commutation spike current; It is the back electromotive force; The armature time constant is determined by the motor structure.

[0121] Therefore, the commutation peak current is: ; in, k =0.01, defined as the commutation overlap angle correction coefficient; That is the rated current.

[0122] Secondly, the multiple of the protection threshold. Defined as: ; in, This is the maximum value of normal current load fluctuation. It is the rated current. It is the no-load current. It is the motor torque with the greatest environmental frictional resistance. It is the motor torque with the lowest environmental frictional resistance. It is the rated load torque.

[0123] (2) Determine whether the motor rotation direction has changed.

[0124] The direction of motor rotation is controlled by the level of two motor control pins. The change in the direction of motor rotation is indicated by the change in the high or low level of the motor control pins. During motor operation, the high and low level signals of the two motor control pins are monitored in real time. If the voltage level signal of the motor control pin remains unchanged, the protection threshold is: ; If the monitoring detects a change in the voltage level of the motor control pin from IN1 high and IN2 low to IN1 low and IN2 high, it indicates a change in the motor's rotation direction. In this case, the protection threshold should be raised to the original protection threshold. of times, that is, updated to .

[0125] (3) Reversing spike shielding.

[0126] The motor is controlled to start rotating from the initial intermediate position, and the rotating current is obtained.

[0127] When the motor reverses direction, the current rotational current is compared with the updated protection threshold. Compare them.

[0128] If the current rotating current is greater than or equal to the protection threshold When this occurs, it indicates that the motor may be stalled or has reached a physical limit fault. At this time, the duration judgment logic is executed, and the stall or physical limit judgment logic and execution measures are further executed.

[0129] After the motor reverses direction, the updated protection threshold is maintained for a set time (e.g., 0.5s) before returning to the original protection threshold.

[0130] (4) Distinguishing between motor stall and physical limit switch and protection.

[0131] If the direction has not reversed, compare the current rotating current with the original protection threshold. Compare them.

[0132] If the current rotating current continues for a set time period (e.g., 1 second) and is greater than or equal to the protection threshold... Then determine the current time node of the rotation.

[0133] Calculate the current rotation time t respectively 当前 With clockwise limit rotation time T 顺 and counterclockwise limit rotation time T 逆 The difference, T 顺 =T 逆 =T / 2.

[0134] Specifically: Calculate the current rotation time t when rotating clockwise. 当前 The remaining time until the clockwise limit is reached, i.e.: Δt 顺 =T 顺 -t 当前 ; When rotating counterclockwise, calculate the current rotation time t. 当前 The remaining time until the counterclockwise limit is reached, i.e.: Δt 逆 =T 逆 -t 当前 ; If Δt 顺 ≤The set time threshold (e.g., 0.5s) indicates clockwise rotation, approaching the clockwise limit, or Δt 逆 If the time threshold is less than or equal to the set time threshold (e.g., 0.5s), it indicates that the rotation is counterclockwise and is close to the counterclockwise limit, then it is marked as a potential limit stop. If Δt 顺 >Set time threshold (e.g., 0.5s) and Δt 逆 If a set time threshold (e.g., 0.5s) indicates that the rotation is in the middle range and there is no rigid limit, it will be marked as a suspected stall.

[0135] After each current over-limit detection, record the rotation direction + the current rotation time t. 当前+ Exceeding limits type (suspected of being stopped / blocked), forming a historical database; statistically analyze the exceeding limits records in the same direction in the historical database: If the frequency of suspected limit switch failures is greater than or equal to the set maximum number of times (e.g., 8 times, meaning that more than 80% of the over-limit occurs within 0.5s of the limit switch), then based on Δt and the historical database, it is determined to be a physical limit switch failure. If the suspected stalling occurs in the historical database (which stores the 10 most recent records) in the same direction and at the same rotation position, with a frequency ≤ the set minimum number of occurrences (e.g., 2 times, including 0 times), then it is determined to be a stalling fault.

[0136] Stall failures generally do not occur at a fixed location, and occur at a low frequency of ≤2 times in 10 samples (accounting for ≤20%), which is consistent with the characteristics of random failures.

[0137] Reasonableness of 0 times: The current over-limit position is a newly appeared random fault point that has never occurred in the historical record. It is a typical scenario of stall fault and fully meets the judgment conditions.

[0138] (5) Differentiated protection.

[0139] If the fault is determined to be a physical limit switch failure, a stop command will be sent immediately, because the limit switch is a rigid block and reversing may cause deformation of the drive shaft, and the indicator light will turn green.

[0140] If a stall fault is detected, a stop command is first sent (lasting 0.5s), followed by a reverse rotation command for 0.3s. After reversing, the current is resampled. If the current returns to normal, the inventory continues. If the current still exceeds the limit, a stop command is sent, and the red light flashes.

[0141] In this embodiment, the multi-level omnidirectional inventory of power materials also includes a step of accurately locating tag data, specifically including: Step S211: Preprocess the tag data using Gaussian filtering. Process the received tag data by weighted averaging. For tag data in a metallic environment, calculate the template coefficient based on the standard deviation. The weight of pixels farther from the center is smaller, forming a weighted window with a smooth transition. Perform sliding convolution operation on the generated Gaussian template on the signal data sequence. Effectively filter out random noise and spike interference caused by metal reflection by weighted averaging. Step S212: Based on the preprocessed tag data, the positioning results of power materials are obtained using a pre-trained positioning model. The positioning model is trained using an optimizer and the error between the predicted coordinate position and the actual position is calculated using the cross-entropy loss function.

[0142] Step S213: Tag data includes signal strength, location coordinates, and timestamp.

[0143] In step S211, because the RSSI location fingerprint data in the database is inevitably affected by non-line-of-sight and multipath effects caused by environmental factors such as interference from metallic materials during the original measurement and collection process, the received RSSI values ​​may have some deviations, affecting the fitting function of the data relationship, thus causing unpredictable errors in the positioning results and having an unforeseen impact on positioning. To improve the reliability and validity of the experimental results, the RSSI experimental data should first be filtered to remove some error values ​​with large deviations, leaving as much effective data as possible.

[0144] This embodiment utilizes the Gaussian filtering principle for processing. In this principle, the RSS value approximately follows the mean... The standard deviation is The probability density function is f ( RSS The Gaussian normal distribution is .

[0145] The Gaussian filter filters out the low-probability RSS values ​​from the Gaussian distribution, and then averages the remaining RSS values ​​to obtain the preprocessed RSS value. The expression for Gaussian filtering can be written as: ; Gaussian filtering can be viewed as a smoothing method that effectively reduces noise interference and signal jitter. In fields such as wireless positioning and indoor navigation, Gaussian filtering is widely used for preprocessing and optimizing RSS positioning. By preprocessing the received signal, the reliability of the signal can be improved, thereby enhancing positioning accuracy and robustness. The mean and standard deviation in Gaussian filtering can be adjusted and optimized according to specific scenarios and application requirements.

[0146] Metal materials in electrical equipment reflect and absorb electromagnetic waves, causing severe attenuation and distortion of tag reading signals. Using a Gaussian filter as a linear smoothing filter can effectively suppress noise introduced by such environmental interference. Optionally, the Gaussian filter processes the signal data received by the RFID reader using a weighted average method. For RFID data in metallic environments: based on the standard deviation... The template coefficients are calculated, with pixels farther from the center having smaller weights, forming a weighted window with a smooth transition. The generated Gaussian template is then subjected to a sliding convolution operation on the signal data sequence. Through weighted averaging, random noise and spike interference caused by metal reflection are effectively filtered out.

[0147] For tags corresponding to non-metallic electrical materials, signal attenuation mainly comes from environmental noise and electromagnetic interference; Gaussian filters are needed to suppress random noise and spike interference. For tags corresponding to metallic electrical materials, the metal reflects the signal, causing severe attenuation; Gaussian filters are needed to compensate for signal loss. Based on this, the standard deviation of the Gaussian filter for tags corresponding to non-metallic electrical materials... The window size can be set based on the normal ambient noise level, with typical parameters being a 3×3 or 5×5 window. The value should be moderate; for tags corresponding to metallic electrical materials, the filter window size needs to be increased (e.g., 7×7) and the value reduced. The value is adjusted to enhance the smoothing effect, while the parameters need to be dynamically adjusted to deal with signal distortion caused by metal reflection.

[0148] Of course, the window size of the Gaussian filter can also be dynamically adjusted based on the intensity of the metal reflection signal. When a strong metal reflection signal is detected, the window size can be increased (e.g., from 3×3 to 5×5 or 7×7) to enhance the smoothing effect; conversely, the window size can be decreased to retain more detail. Increasing the window size can more effectively suppress high-frequency noise, but may result in the loss of some signal detail, requiring a balance based on the actual scenario.

[0149] In some embodiments, the signal-to-noise ratio (SNR) of the signal can also be dynamically adjusted by real-time monitoring. Value. If signal distortion is severe (e.g., low SNR), reduce. The value can be decreased (e.g., from 1.0 to 0.5) to enhance the smoothness of the filter; if the signal quality is good, it can be increased appropriately. This is to retain more high-frequency components. The smaller the value, the flatter the Gaussian distribution and the smoother the filtering effect, but it may overly blur the effective signal.

[0150] In some embodiments, a supplementary scanning step is also included, which specifically includes: determining the frequency of missed scanning angles for each power material based on the existing power material ID, material type, storage location, initial missed scanning angle, supplementary scanning angle range, offset threshold used, supplementary scanning speed, supplementary scanning result, and scanning time, thereby determining the easily missed scanning areas of the power materials, and performing initial supplementary scanning on the easily missed scanning areas.

[0151] Specifically, a structured data table can be designed with fields including power material ID, power material type (metal / non-metal), storage location (near metal shelves / open area), initial missed scan angle, supplementary scan angle range, offset threshold used, supplementary scan speed, supplementary scan result (detected / not detected), and scan time, to build a missed scan-scanning database.

[0152] It can also classify and statistically analyze the data in the missed scan-scan database by power material type and storage location, calculate the frequency of missed scan angles for different shelf types, different power material storage areas, and different power materials, generate exclusive labels for easily missed scan areas of power materials, and update the initial scan priority list of the inventory equipment.

[0153] In this embodiment, considering the strong directivity of ultra-high frequency signals, the initial missed scan angle directly corresponds to the multipath interference intensity. Based on the logic that the interference intensity is positively correlated with the threshold requirement, a fine-tuning rule is formulated: Initial missed scan angle is 0°-30° (close to the front, strong interference): signal energy is concentrated but metal reflection interference is the most serious. Take the upper limit of the basic threshold (25° for metal, 20° for non-metal) to ensure coverage of all potential missed scan areas.

[0154] Initial missed scan angle is 30°-60° (side front, moderate interference): signal energy and interference are reduced synchronously, and the middle limit of the basic threshold is taken (20° for metal, 15° for non-metal) to balance coverage and efficiency.

[0155] Initial missed scan angle is 60°-90° (lateral, weak interference): Although the signal energy is low, the interference is extremely weak. Take the lower limit of the basic threshold (15° for metals, 10° for non-metals) to avoid excessive expansion of the scanning range and resulting in wasted efficiency.

[0156] This is because UHF signals have strong directionality. Within the range of 0°-30° directly in front, the signal energy is concentrated, but interference from metal reflections is also most concentrated. Beyond 30°, the signal energy gradually attenuates, but interference also decreases accordingly. Above 60°, although the signal energy is low, the interference is extremely weak, making it easier to receive stably (requiring only a small-scale rescan). Therefore, the offset threshold angle range is consistent with the signal's own propagation attenuation characteristics.

[0157] In some embodiments, the scanning speed can be adaptively adjusted by combining the offset threshold and the type of power material, clarifying the purpose of the speed reduction, increasing the residence time of the UHF signal in the target area that is prone to being missed, offsetting the signal attenuation caused by multipath effect, improving the reading success rate, and avoiding excessive speed reduction that leads to a decrease in efficiency.

[0158] When adjusting the speed, it's important to note that signal dwell time is inversely proportional to scanning speed: Dwell time (seconds / degree) = 1 ÷ Scanning speed (degrees / second). For example, at a speed of 5° / s, the dwell time is 0.2 seconds per degree; at a speed of 2° / s, the dwell time is 0.5 seconds per degree. A longer dwell time enhances the ability to counteract multipath effects and capture weak signals.

[0159] The standard scanning speed (5° / s) is the industry benchmark, while the speed classification corresponding to the offset threshold is derived through dual derivation of scanning range calculation and dwell time verification. The core is to ensure that the total rescanning time is reasonable under different offset thresholds, and that the dwell time is sufficient to cover interference.

[0160] Of course, in actual implementation, optimization algorithms, such as particle swarm optimization or gray wolf optimization, can be selected to optimize the solution and determine a better fine-tuning scheme or speed adjustment scheme, which will not be elaborated here.

[0161] In this embodiment, as Figure 4 As shown, the process of the digital twin coupling verification step for power materials in step S3 includes: Step S301: Obtain image data of the storage area; Step S302: Process the image data using the trained image detection model to obtain the identification results of power materials; Step S303: Using the digital twin model and the recognition results, determine the spatial coordinates of the identified power materials in the digital twin model (which can correspond to the actual spatial coordinates) and associate them with the corresponding storage locations; Step S304: Retrieve the business knowledge base data corresponding to the current inventory task. The business knowledge base data includes at least the theoretical inventory list of the target warehouse area. Step S305: Couple and verify the identification results, actual spatial coordinates, and associated cargo location information with the business knowledge base data to generate verification results.

[0162] This embodiment uses image acquisition by a drone as an example for illustration.

[0163] The flight path of the drone in the warehouse area to be inventoried for electrical materials is pre-planned. The flight path covers the storage areas of all target locations within the warehouse, and preset parameters for flight altitude, speed, and shooting interval are set to ensure that the drone's onboard camera can completely capture image data of the materials in each location. The drone executes its flight mission according to the preset flight path, capturing continuous image frames of the area to be inventoried in real time through its onboard visual sensor. Simultaneously, the drone's pose estimation system acquires its own three-dimensional pose data in a preset coordinate system in real time, including position coordinates and attitude parameters. The drone synchronously associates the real-time image frame data with the corresponding three-dimensional pose data through timestamp matching, ensuring that each image frame corresponds to a unique real-time pose data of the drone. Subsequently, the associated image data and three-dimensional pose data are continuously uploaded to an edge server or cloud according to a preset transmission strategy.

[0164] The image data and pose data are input into a trained 3D coupled verification model to generate and output verification results; wherein, the 3D coupled verification model processes the image data through a visual recognition unit to identify and obtain identification information of power materials.

[0165] After receiving the image data to be processed uploaded by the UAV, the visual recognition unit first performs image enhancement operations, including adjusting the image brightness and contrast based on gamma correction, removing image noise through Gaussian filtering, and optimizing image details using histogram equalization to ensure that the image meets the quality requirements of subsequent processing.

[0166] Existing object detection models can be used for training. The trained model is then used to detect preprocessed images, outputting the predicted bounding box coordinates, material type, and detection confidence of the nameplates of power materials in the image. Valid prediction results are filtered by setting a pre-set confidence threshold, and redundant bounding boxes are eliminated through non-maximum suppression processing to determine the unique nameplate region. This region is then used for identification to obtain complete power material identification information. The identification information includes several of the following: material type, model, specifications, capacity, factory number, identification confidence, and ROI of the nameplate image. Finally, the power material identification information is used as a unique material identifier.

[0167] The identification process and identification model can be selected from existing models, and will not be elaborated here.

[0168] The system retrieves preset global coordinate system parameters from the digital twin spatial base and simultaneously extracts the drone's own coordinate system parameters carried in the drone's real-time pose data. A coordinate transformation matrix is ​​then used to convert the drone's position coordinates and attitude parameters in its real-time pose data into pose data in the warehouse's global coordinate system, thus aligning the real-time pose data with the digital twin spatial base coordinate system.

[0169] The intrinsic parameter matrix of the UAV's onboard camera is obtained, and combined with the aligned global pose data of the UAV, a camera imaging model is established. From the pixel coordinates of the power material nameplate area output by the visual recognition unit, the center pixel of the nameplate is selected as the material feature point.

[0170] Based on the principle of perspective projection, the actual spatial coordinates of the identified power materials in the digital twin spatial base are calculated. The location space data interface of the digital twin spatial base is called to obtain the predefined bounding box parameters of all target locations within the warehouse. The bounding box of each location corresponds to the minimum and maximum three-dimensional coordinates in the global coordinate system of the warehouse. An inclusion check is performed on the actual spatial coordinates of the identified power materials and the bounding box parameters of each location. If there exists one and only one location i that satisfies the above spatial inclusion condition, the unique location D of that location is extracted and associated with the unique identifier of the identified power materials, generating material-location association data, which is synchronously stored in the material location mapping table of the digital twin spatial base and the inventory association table of the business knowledge base.

[0171] The system retrieves basic information about the current power material inventory task from the management platform. This basic information includes at least a unique task identifier, target warehouse area number, inventory time range, and task priority. The system then matches the corresponding task configuration record in the "Inventory Task Index Table" of the business knowledge base using the unique task identifier, and extracts the spatial range parameters of the target warehouse area and the associated inventory data query permissions.

[0172] Based on the parsed target warehouse area number and combined with the warehouse area-inventory association rules in the business knowledge base, the query dimensions for the theoretical inventory list are determined, and these dimensions are integrated into a structured query statement. By initiating a query request to the inventory database of the knowledge base, the business knowledge base extracts records that meet the query conditions from the WMS theoretical inventory master table. Each record contains at least the location ID, the expected material unique identifier, the expected material quantity, the most recent inventory count time, and the material lifecycle status, forming a structured theoretical inventory list for the target warehouse area. Finally, the standardized theoretical inventory list is associated and bound with the unique identifier of the current inventory count task, completing the entire retrieval process.

[0173] The coupling verification in this embodiment is a three-dimensional coupling verification based on vision-space-business information, specifically including: Spatial verification determines whether the actual spatial coordinates of the electrical materials fall within the bounding box of their expected storage location. Specifically, the expected bounding box parameters and maximum three-dimensional coordinates of the associated storage location are retrieved from the digital twin spatial base. The actual spatial coordinates of the identified electrical materials are then compared with the expected bounding box parameters to determine whether the spatial inclusion condition is met. If not, an anomaly is marked as "spatial coordinates deviate from expected storage location".

[0174] The business verification process determines whether the identity and / or quantity of the identified power materials match the expected material information for the corresponding storage location in the theoretical inventory list. Specifically, it extracts the expected material information corresponding to the associated storage location from the theoretical inventory list in the business knowledge base. This expected material information includes at least a unique identifier for the expected material and the expected quantity. The power material identification information output by the visual recognition unit is compared with the expected material information. If the unique identifiers are inconsistent, an "identity mismatch" anomaly is marked; if the unique identifiers are consistent but the quantities are mismatched, an "quantity mismatch" anomaly is marked.

[0175] Status verification, based on the business rule base in the business knowledge base, determines whether the existence status of the identified power materials conforms to predefined business rules. Specifically, it retrieves predefined business rules from the business rule base in the business knowledge base. These business rules include at least material lifecycle status rules and material storage association rules. Based on these business rules, the existence status of the identified power materials is verified. If the material lifecycle status is "scrapped" or "pending inspection" but appears within the normal inventory range, it is marked as "material status violation". If the material storage violates the association rules, it is marked as "storage association violation". Finally, the results of spatial verification, business verification, and status verification are summarized to generate a structured verification result containing the location ID, material ID, verification result, verification confidence level, timestamp, and anomaly type. The verification result is then packaged according to a preset format and synchronously pushed to the digital twin screen and verification result database of the management platform to complete the entire process of three-dimensional coupled verification.

[0176] like Figure 5 As shown, step S4 includes: Step S401: Obtain data on the physical state, environmental impact, operating load, and life cycle of power materials, perform preprocessing, fill in missing data using a pre-trained neural network model, eliminate dimensional differences using Z-score standardization, and filter out abnormal data. Step S402: According to the category of power materials, extract the features of the preprocessed multi-source data. Among them, for sudden materials, extract several of the following: oil level change rate, oil temperature rise rate, insulation resistance decay rate, and oil leakage image confidence. Several factors were extracted from progressively aging materials, including partial discharge signal amplitude, insulation aging index, joint temperature deviation, and ambient temperature difference. For environmentally sensitive materials, extract several of the following: salt spray concentration, humidity product, surface contamination level, and flashover voltage threshold deviation; Step S403: Employ a multi-head attention mechanism to dynamically allocate the weights of the extracted features and perform weighted fusion to obtain the fused features; Step S404: Utilize Bayesian network modeling to model the correlation between features and faults. Based on the fused features, update the probability of risk occurrence. Use the analytic hierarchy process (AHP) to determine the severity of the consequences of fault types. Combine the probability of risk occurrence and the severity of consequences to determine the risk level. Use reinforcement learning algorithms to determine the scheduling scenarios and related material measures for power supplies.

[0177] In this embodiment, in step S401, as shown in Table 1, four-dimensional core data is collected based on the safety characteristics of different power materials. When acquiring the data, it is necessary to ensure real-time performance and completeness.

[0178] Table 1

[0179] In this embodiment, LSTM is used to fill in missing data, Z-score standardization is used to eliminate dimensional differences, edge nodes filter abnormal data (such as sensor failures) in real time, and global data is stored in the cloud for model training.

[0180] Step S402, material customization feature extraction, specifically: 1) High-consequence emergency materials (transformers): Core characteristics = oil level change rate (dL / dt), oil temperature rise rate (dT / dt), insulation resistance decay rate, and oil leakage image confidence level; 2) Gradual aging materials (cables): Core characteristics = partial discharge signal amplitude, insulation aging index (based on years of operation + load), joint temperature deviation (difference from ambient temperature); 3) Environmentally sensitive materials (insulators): Core characteristics = salt spray concentration × humidity product, surface contamination level, flashover voltage threshold deviation.

[0181] In this embodiment, step S403 involves dynamically allocating feature weights using a multi-head attention mechanism, for example: When a transformer oil leakage warning is issued, the attention weight (0.6) of the oil level change rate is much higher than the ambient temperature weight (0.1). When the insulator is in a coastal area, the salt spray concentration weight (0.5) is higher than the service life weight (0.2).

[0182] The fusion formula is: ; For attention weights, It is a single feature value.

[0183] In this embodiment, in the risk assessment in step S404, a three-dimensional risk assessment system is formed by calculating the probability of risk occurrence (P), the severity of consequences (C), and the dynamic risk entropy (E).

[0184] First, the probability of risk occurrence (P) is calculated: In this embodiment, the correlation between features and faults is modeled based on a Bayesian network (BN), and the fused feature F is input to update the posterior probability: P(Fault|F)=P(F|Fault)P(Fault) / P(F); Where P(Fault) is the prior probability of a historical fault, such as the prior probability of oil leakage of a certain type of transformer being 0.05, and P(F|Fault) is the conditional probability of feature F when the fault occurs, which is obtained through training with historical data.

[0185] Calculation of the severity of consequences (C): In this embodiment, the analytic hierarchy process (AHP) is used to construct the indicator system, and the weights are optimized through expert scoring and data verification, as shown in Table 2.

[0186] Table 2

[0187] ; in, As weight, For the dimension score, the value of C ranges from [0, 1].

[0188] Dynamic risk entropy (E) calculation: Describing the volatility uncertainty of P and C (the higher the entropy, the more likely the risk is to escalate suddenly):

[0189] The dynamic risk entropy ranges from [0, 2], and when E ≥ 0.6, it represents high volatility risk.

[0190] The risk classification is shown in Table 3.

[0191] Table 3

[0192] Management and decision-making level: Strengthen learning-driven dynamic management and control In this embodiment, based on risk classification, deep reinforcement learning (DRL) is used to optimize the control strategy, achieving the optimal balance between safety and cost.

[0193] The deep reinforcement learning model is designed as follows: State: S = (R, E, Resource, Status), where Resource is the currently available controllable resources (inspection personnel, maintenance equipment), and Status is the material operation status (running / shutdown). Action: A set of control measures (e.g., real-time monitoring, increasing inspection frequency, shutdown for maintenance, component replacement, load optimization). Reward function: ,in: SafeGain is the risk reduction amount, and its value is R. before -R after That is, the difference between the risk value R before and after; Cost: Control costs (personnel + equipment + power outage losses, standardized); λ is the safety weight (λ=0.9 for extremely / high risk, λ=0.7 for medium risk, and λ=0.5 for low risk).

[0194] Examples of customized management and control strategies are shown in Table 4: Table 4

[0195] When the risk level reaches extreme danger or a sudden failure occurs, emergency response is automatically triggered, which is achieved through risk diffusion simulation and resource optimization scheduling.

[0196] Risk diffusion simulation: In this embodiment, a power material topology model is constructed based on digital twin (DT), and graph neural network (GNN) is used to model the relationships between materials (such as the connection between transformer-cable-switch cabinet). The GNN takes faulty nodes (such as transformer explosion early warning) as input and outputs risk propagation paths and impact ranges (such as the probability of fire spreading to surrounding cables and the area affected by power outages).

[0197] Emergency resource allocation: Based on the risk diffusion results, an improved genetic algorithm (GA) is used to optimize resource allocation: Objective function: min(synchronization path + total scheduling time + scheduling energy consumption); Constraints: resource capacity limitations, road access conditions, power supply priority (e.g., hospitals > residential users).

[0198] Specifically, acquire material data, including the type of power material (e.g., cable connectors, small insulators, maintenance wrenches), weight (e.g., 5kg, 3kg, 8kg), volume, characteristics (e.g., precision / normal), storage slot number, and location / distance of the target dispatch area.

[0199] Acquire mobile device data, such as the parameters of two transport AGVs (AGV1 and AGV2) in this embodiment: driving speed v=1.2m / s, turning radius of 0.3m, unit driving energy consumption k_drive=0.0004kWh / m, unit standby energy consumption k_standby=0.0001kWh / s, maximum load capacity of 10kg, and compatible material types.

[0200] It is also necessary to collect the channel congestion index (0-1, 0 for smooth traffic and 1 for congestion), the location of temporary obstacles, the coordinates of the personnel work area, and the real-time location and remaining battery power of the mobile AGV. Real-time reception of dispatch tasks (e.g., Task 1: transporting 5kg cable joints from area A to area B; Task 2: transporting 3kg insulators from area B to area C, as well as the loading and unloading time of the corresponding materials, etc.).

[0201] When receiving dispatch instructions, if it is an emergency repair instruction (such as material requisition for power outage repair), the task priority is marked as "urgent". If it is a regular dispatch, it can be marked as a daily material requisition / replenishment task, load rate 30%-60%, working hours (9:00-18:00); if it is an off-peak dispatch, it can be marked as a load rate <30%, non-working hours (22:00-6:00 the next day), no personnel working; if it is a high-load dispatch, it can be marked as a load rate >60%, concentrated outbound periods, and congestion index.

[0202] The acquired data is then cleaned: sensor anomalies (such as equipment location data outside the warehouse area) are removed; after unifying the Cartesian coordinate system with the inbound / target area as the origin, the data is then processed according to the scheduling scenario: the scenario type and dynamic constraint parameters are associated, such as amplifying the congestion index weight during high-load scheduling and removing the passage restrictions of non-core channels during emergency scheduling.

[0203] A scenario-based multi-objective function F = w_L×L_total + w_T×T_total + w_E×E_total (where F is minimized is the optimization direction) is constructed, and the weights are dynamically adjusted according to the scenario to ensure that core requirements are prioritized. The definitions, weight allocations, and calculation methods of each parameter are shown in Table 5.

[0204] Table 5

[0205] The weights satisfy w_L + w_T + w_E = 1.

[0206] For specific settings, please refer to Table 6.

[0207] Table 6

[0208] In some embodiments, a safety objective also needs to be added. This objective is not included in the weight calculation but is used as a prerequisite constraint. For example, if the safety coefficient S_score of a certain path scheme is less than 60, the scheme will be directly eliminated regardless of how small the F value is. If the safety coefficient S_score of multiple schemes is greater than or equal to 60, the optimal scheme is selected by comparing the F-values.

[0209] Core rule: All scenarios must meet S_score≥60 (basic safety threshold), and high-load scheduling additionally requires S1 (personnel distance score)≥20 points (i.e., safe distance ≥0.8m); Safety factor calculation: S_score = S1 (personnel distance) + S2 (equipment distance) + S3 (adaptation) + S4 (speed limit) (25 points for each item).

[0210] Emergency response plan generation: Automatically output standardized emergency plans, such as "cut off the 10kV side power supply of the faulty transformer → evacuate personnel within 50 meters → dispatch the nearest emergency repair team (within 3km) → prioritize restoring power supply to the hospital → replace the leaking transformer".

[0211] In some embodiments, self-learning updates can also be incorporated to continuously optimize model performance algorithms and achieve self-evolution through closed-loop data feedback: After fault handling, update the prior probability P(Fault) and conditional probability P(F|Fault) of the Bayesian network. Based on actual control effectiveness (such as risk reduction rate and cost consumption), optimize the reward function and attention mechanism weights of RL; Accumulate material failure data from different regions and environments to form a knowledge base linking "materials-environment-failure" and improve the accuracy of risk prediction.

[0212] Example 2 A multi-objective collaborative intelligent management and control system for power materials under multiple scheduling scenarios includes: The warehouse location precision allocation unit is used to determine the type and quantity of required power materials according to the needs of various scheduling scenarios, generate demand instructions, and determine the storage priority of power materials under storage constraints based on the warehousing impact coefficient of the power materials to be allocated and the scheduling scenario type, thereby determining the warehousing location allocation strategy. The multi-level omnidirectional inventory unit for power materials is used to determine the adaptive sorting factor based on the weighted sum of the signal strength, phase stability, polarization matching degree and missed reading compensation factor of the power materials tags in the warehouse, and to perform omnidirectional hierarchical inventory based on the priority of the adaptive sorting factor. The digital twin coupling verification unit for power materials is used to identify the semantic information of power materials based on the image data of the acquired storage area, determine the spatial coordinates of the corresponding power materials in the storage digital twin model, associate them with the corresponding storage location, and perform coupling verification by combining the identified semantic information, spatial coordinates and storage location allocation strategy to update the storage management information. The multi-objective collaborative optimization scheduling unit is used to comprehensively consider the physical state, environmental impact, historical operating load and life cycle data of power materials, conduct risk assessment, predict the scheduling scenarios of power materials, and combine the storage management information of each storage area to achieve multi-objective collaborative optimization with the goal of achieving the shortest path, the shortest total scheduling time and the lowest scheduling energy consumption, and pre-generate scheduling strategies.

[0213] Example 3 A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the steps in the method provided in Embodiment 1.

[0214] Example 4 An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the steps in the method provided in Embodiment 1.

[0215] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-target collaborative intelligent management and control method for electric power materials in a multi-scheduling scenario, characterized by, The method comprises the following steps: The warehouse location precise allocation step: according to the requirements of various scheduling scenarios, the types and quantities of required power materials are determined, demand instructions are generated, the storage priority of power materials is determined under the storage condition constraint according to the warehouse influence coefficient of the power materials to be allocated combined with the type of the scheduling scenario, and then the warehouse location allocation strategy is determined; The power material multi-level omnidirectional inventorying step: according to the weighted sum of the signal strength, phase stability, polarization matching degree and missed reading compensation factor of the warehouse power material label, the adaptive sorting factor is determined, and omnidirectional hierarchical inventorying is performed according to the priority of the adaptive sorting factor; The power material digital twin coupling verification step: according to the obtained image data of the warehouse area, the semantic information of the power material is recognized, the spatial coordinates of the corresponding power material in the warehouse digital twin model are determined, and are associated to the corresponding warehouse location. Coupling verification is performed combined with the recognized semantic information, spatial coordinates and warehouse location allocation strategy, and the warehouse management information is updated; The multi-objective collaborative optimization scheduling step: considering the characteristics of the physical state, environmental influence, historical operation load and life cycle data of the power material, risk assessment is performed, the scheduling scenario of the power material is predicted, and the warehouse management information of each warehouse area is combined to achieve the multi-objective collaborative optimization of the shortest path, the shortest total scheduling time and the lowest scheduling energy consumption, and the scheduling strategy is pre-generated.

2. The multi-target collaborative intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 1, characterized in that, The process of determining the adaptive sorting factor according to the weighted sum of the signal strength, phase stability, polarization matching degree and missed reading compensation factor of the warehouse power material label comprises: obtaining multiple label data in the warehouse area; The adaptive sorting factor of the multiple label data is obtained through the weighted sum of the signal strength, phase stability, polarization matching degree and missed reading compensation factor, and the weights corresponding to the signal strength, phase stability, polarization matching degree and missed reading compensation factor are adjusted according to the different types of power equipment in the power warehouse; The label data with an adaptive sorting factor value higher than a preset value is scanned preferentially, and then other label data is scanned according to the adaptive sorting factor.

3. The multi-target collaborative intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 2, characterized in that, Adaptive ranking factor For: ； wherein, is a signal strength criterion value; is a phase stability criterion value; is a polarization matching criterion value; is a missed read compensation factor criterion value; , , and are weights corresponding to signal strength, phase stability, polarization matching, and missed read compensation factor, respectively. The label data of the equipment with a missed reading compensation factor standard value greater than a preset value and not read successfully for the first time is subjected to an additional retry round.

4. The multi-target collaborative intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 2, characterized in that, The signal strength standard value is obtained by proportionally converting the minimum value and the maximum value of the readable signal to 0-1, and the closer to the maximum value, the closer to 1; the stability or change rate of the label signal phase is determined, and the stability or change rate is compressed to 0-1 through a function to obtain the phase stability standard value; the closer the phase is to the stable phase, the higher the score is; the matching degree of the label direction and the antenna polarization direction is determined, and the matching degree is mapped to 0-1 to obtain the polarization matching degree standard value; the closer the direction is to the consistent direction, the higher the score is; The missed reading compensation factor standard value is determined in 0-1 by counting the number of the most recent unreadings of the label, and the more the missed readings, the closer to 1.

5. The multi-target collaborative intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 1, characterized in that, The omnidirectional hierarchical inventorying is performed by using the omnidirectional inventorying device, specifically comprising: according to the rotation current of the motor of the omnidirectional inventorying device at a set rotation speed, the forward and reverse limiting rotation time is matched by using the constructed rotation speed-time-current correspondence table; A friction compensation coefficient is defined according to the dust concentration and the forward and reverse limiting rotation angles, a time compensation amount is determined in combination with the cumulative running time of the motor, and a final rotation time is obtained according to the forward and reverse limiting rotation time and the time compensation amount; The position of the motor after rotating in the opposite direction by T / 2 from any physical limit is taken as an initial intermediate position, T is the final rotation time, the motor is controlled to rotate from the initial intermediate position, and the rotating current is obtained; When the motor commutates, the protection threshold is updated according to the commutation peak current, and the current rotating current is compared with the updated protection threshold; if the motor does not commutate, the current rotating current is compared with the original protection threshold, so as to determine whether a fault occurs, and the fault type is determined according to the time node at which the current rotation time is located when the fault occurs.

6. The multi-target coordinated intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 5, characterized in that, The rotation speed-time-current correspondence table is constructed according to the forward and reverse limiting rotation time and the rotating current when the motor rotates from the clockwise physical limit to the counterclockwise physical limit at different rotation speeds; The friction compensation coefficient is ; The time compensation amount is ; where t is the cumulative running time; is the forward / reverse limiting rotation angle, is the rotation speed in the clean state, is the rotation speed difference in the actual state, C is the dust concentration, is a proportional constant.

7. The multi-target coordinated intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 5, characterized in that the original The protection threshold is defined as a multiple of the rated current, and the updated protection threshold is defined as a multiple of the commutation peak current; wherein the multiple is: ; wherein, is the maximum normal current load fluctuation, is the rated current, is the no-load current, is the maximum motor torque for environmental frictional resistance, is the minimum motor torque for environmental frictional resistance, is the rated load torque, is the current variation amount; The commutation spike current is: ; wherein, k is a commutation overlap angle correction factor, is a rated voltage; is a back EMF; is an armature inductance, is a time interval.

8. The multi-target collaborative intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 5, characterized in that, The process of determining the fault type according to the time node at which the current rotation time is located includes: If the current rotating current lasts for a set time period and is greater than or equal to the original protection threshold, the time node at which the current rotation time is located is determined; Calculate the current rotation time t respectively 当前 Remaining time Δt until clockwise limit is reached 顺 The remaining time Δt from the counterclockwise limit 逆 If Δt 顺 ≤ Set time threshold or Δt 逆 If Δt is less than or equal to the time threshold, it is marked as a potential limit switch; if Δt 顺 >Time threshold and Δt 逆 When the time threshold is reached, it is marked as a suspected stalling situation; If the frequency of the limit suspicion is greater than or equal to the set maximum number threshold, it is determined that a physical limit fault occurs; if the frequency of the stall suspicion at the same rotation position in the same direction is less than or equal to the set minimum number threshold, it is determined that a stall fault occurs.

9. The multi-target collaborative intelligent management and control method for electric power materials in a multi-scheduling scenario according to claim 1, characterized in that, The power material multi-level omnidirectional inventorying step further includes a label data accurate positioning step, specifically including: Gaussian filtering is used for pre-processing of the label data, the received label data is processed by weighted average, for label data in a metal environment, a template coefficient is calculated according to a standard deviation, the weight of a pixel farther from the center is smaller, a weighted window with smooth transition is formed, a generated Gaussian template is subjected to sliding convolution operation on a signal data sequence, and random noise and peak interference caused by metal reflection are effectively filtered out by weighted average; According to the pre-processed label data, a positioning result of the power material is obtained by using a pre-trained positioning model, the positioning model is trained by using an optimizer, and a cross-entropy loss function is used to calculate the error between the predicted coordinate position and the real position. The label data includes signal strength, position coordinates and time stamp.

10. The method for multi-objective collaborative intelligent management and control of power materials in a multi-scheduling scenario as described in claim 9, characterized in that, For labels corresponding to metal power materials, the filter window size is increased and the standard deviation is reduced to enhance the smoothing effect, and the parameters are dynamically adjusted according to the signal distortion caused by metal reflection; as the metal reflection signal strength increases, the window size is increased; otherwise, the window size is reduced to retain more details. The Gaussian filtering expression is: ； wherein is the standard deviation; is the value approximately obeys the mean.

11. The method for multi-objective collaborative intelligent management and control of power materials in a multi-scheduling scenario as described in claim 1, characterized in that, The power material multi-level omnidirectional inventorying step further includes a supplementary scanning step, specifically including: determining the missing scanning angle frequency of each power material according to the existing power material ID, material type, storage location, initial missing scanning angle, supplementary scanning angle range, used offset threshold, supplementary scanning speed, supplementary scanning result and scanning time, and then determining the easy-to-miss scanning area of the power material, and performing initial supplementary scanning on the easy-to-miss scanning area.

12. The method of claim 11, wherein the method further comprises: The base offset threshold interval is determined according to the material type, and the offset threshold is determined according to the initial missed scanning angle, wherein the base offset threshold interval of metal materials is set to 15°-25°; the base offset threshold interval of non-metal materials is set to 10°-20°, if the initial missed scanning angle is within 0°-30°, the upper limit of the corresponding base offset threshold interval is taken; if the initial missed scanning angle is within 30°-60°, the middle limit of the base offset threshold interval is taken; if the initial missed scanning angle is within 60°-90°, the lower limit of the base offset threshold interval is taken.

13. The method for multi-objective collaborative intelligent management and control of power materials in a multi-scheduling scenario as described in claim 12, characterized in that, The actual scanning range of the supplementary scanning is determined according to the offset threshold, the scanning speed is determined according to the offset threshold and the material type, and the supplementary scanning in the actual scanning range is controlled, specifically, the actual scanning range of the supplementary scanning is twice the offset threshold, if the offset threshold is greater than 20°, the scanning speed is reduced to 2° / s; if the offset threshold is within 10°-20°, the scanning speed is reduced to 3° / s; if the offset threshold is less than 10°, the scanning speed is reduced to 4° / s; and the scanning speed of metal materials is further reduced by 0.5° / s based on the above scanning speed, and the scanning speed of non-metal materials remains unchanged.

14. The method for multi-objective collaborative intelligent management and control of power materials in a multi-scheduling scenario as described in claim 1, characterized in that, The warehouse location allocation step specifically comprises: determining the required power materials and their current inventory according to the maintenance operation demand and the emergency repair operation demand, and issuing a help demand when the current inventory is lower than the set threshold; An influence coefficient of the warehouse-in of the to-be-allocated power materials is obtained, including weight, volume, warehouse-in and warehouse-out frequency, storage requirement and help demand of other warehouses; The weight of each warehouse-in influence coefficient is determined according to the operation type to obtain the power material storage priority; Under the constraint of meeting the storage requirement, the warehouse location is selected according to the power material storage priority to obtain the warehouse location allocation scheme.

15. The method of claim 14, wherein the method further comprises: If the maintenance operation demand is received, the inventory of the power materials in the current warehouse is checked to determine the required power materials and their current inventory; If the maintenance operation demand is not received, and the current time is in the spring inspection period or the autumn inspection period, the inventory of the power materials in the current warehouse is checked to determine the required power materials and their current inventory; If it is not in the spring inspection period or the autumn inspection period and the maintenance operation demand is not received, only the inventory is checked.

16. The method of claim 14, wherein the method further comprises: The specific process of determining the weight of each warehouse-in influence coefficient according to the operation type to obtain the power material storage priority comprises: the weight, volume, warehouse-in and warehouse-out frequency, storage requirement and help demand of other warehouses are quantitatively processed according to the set grade threshold respectively in advance; The quantification of the storage condition includes ordinary, moisture-proof, constant temperature and humidity and special hazardous goods; The quantification of the help demand of other warehouses includes regular inventory, help demand reservation, 24-hour execution of help demand and help site waiting for power materials; The operation type includes maintenance operation, mixing operation and emergency repair operation, the operation coefficient is determined according to the operation type, and the weight of each warehouse-in influence coefficient is determined as follows: The weight of the warehouse-in and warehouse-out frequency F is γ (k) = γ0+k*(γ1-γ0); The weight of the storage condition C is δ (k) = δ0+k*(δ1-δ0); The weight of the space V is β (k) = 0.1-0.05k; The weight of the help demand of other warehouses E is ε (k) = 0.1+0.05k; The weight a of the weight W is a constant set; Wherein, γ0 and δ0 are the frequency of warehouse in and out and storage condition basic weights of the periodic maintenance operation mode respectively; γ1 and δ1 are the frequency of warehouse in and out and storage condition basic weights of the emergency repair operation mode respectively; The power material storage priority L is: L = a (k) * W + β (k) * V + γ (k) * F + ε (k) * E - δ (k) * C; Under the constraint of meeting the storage requirements, the storage area meeting the storage requirements is screened out, and then the storage location is sorted and selected according to the power material storage priority.

17. The method of claim 1, wherein the method further comprises: determining a power supply target for each of the power supply sources; and determining a power demand target for each of the power demand sources. In the process of the power material digital twin coupling verification step, the image data of the storage area is acquired, and the target detection model is used to locate the nameplate area in the image data; The located nameplate area image is perspective transformation corrected to obtain a standard front view; The corrected nameplate image is subjected to optical character recognition, and based on the fixed field structure of the power material nameplate, the structured key information is extracted; The extracted key information is combined to generate a unique material identifier.

18. The method for multi-objective collaborative intelligent management and control of power materials in a multi-scheduling scenario as described in claim 1, characterized in that, Comprehensively considering the characteristics of the physical state, environmental influence, operation load and life cycle data of the power material, the risk assessment is carried out, and the process of predicting the dispatching scene of the power material includes: acquiring the physical state, environmental influence, operation load and life cycle data of the power material, preprocessing, filling in the missing data by using the pre-trained neural network model, eliminating the dimension difference by using Z-score standardization, and filtering the abnormal data; According to the power material category, the features of the preprocessed multi-source data are extracted, wherein some of the oil level change rate, oil temperature rise rate, insulation resistance decay rate and oil leakage image confidence are extracted for the emergency material; Some of the partial discharge signal amplitude, insulation layer aging index, joint temperature deviation and environmental temperature difference are extracted for the gradual aging material; Some of the salt fog concentration, humidity product, surface contamination degree and flashover voltage threshold deviation are extracted for the environment sensitive material; The multi-head attention mechanism is used to dynamically allocate the weight of the extracted features, and the weighted fusion is carried out to obtain the fusion features; The association between the features and the faults is modeled by using the Bayesian network, the risk occurrence probability is updated according to the fusion features, the consequence severity of the fault type is determined by using the analytic hierarchy process, the risk level is judged by comprehensively considering the risk occurrence probability and the consequence severity, and the dispatching scene and the related material measures of the power material are determined by using the reinforcement learning algorithm.

19. The method of claim 1, wherein the method further comprises: determining a power supply target for each of the power supply sources; and determining a power demand target for each of the power demand sources. The process of pre-generating the dispatching strategy includes: acquiring the power material parameters, positions, number and performance parameters of the mobile devices of each storage area; Acquiring the channel traffic conditions, temporary obstacle positions and personnel operation area coordinates between each storage area and the dispatching target area; Acquiring the real-time position, speed and remaining power of the mobile device; According to the predicted scheduling scene of electric power materials and the related material measures, a scenario-based multi-objective function F = w_L×L_total + w_T×T_total + w_E×E_total is constructed, with the minimum of the target function F as the optimization direction, w_L, w_T and w_E as weights, wherein L_total is the total length of a single task path, T_total is the total scheduling time, the driving time, the loading and unloading time and the waiting time are comprehensively considered, and E_total is the total scheduling energy consumption; The weights are different for different scheduling scenes of electric power materials, and are dynamically adjusted according to the scheduling scenes. 20.A multi-objective collaborative intelligent management and control system for electric power materials under multiple scheduling scenes, characterized in that it comprises: A warehouse location precise allocation unit for determining the types and quantities of required electric power materials according to the requirements of various scheduling scenes, generating a demand instruction, determining the storage priority of electric power materials under the storage condition constraint according to the warehouse influence coefficient of the electric power materials to be allocated combined with the scheduling scene type, and then determining the warehouse location allocation strategy; A multi-level omnidirectional inventory unit for electric power materials, for determining an adaptive sorting factor according to the weighted sum of the signal strength, phase stability, polarization matching degree and missing read compensation factor of the warehouse electric power material label, and performing omnidirectional hierarchical inventory according to the priority of the adaptive sorting factor; An electric power material digital twin coupling verification unit for identifying the semantic information of electric power materials according to the image data of the warehouse area, determining the spatial coordinates of the corresponding electric power materials in the warehouse digital twin model, associating to the corresponding warehouse location, and coupling and verifying with the identified semantic information, spatial coordinates and warehouse location allocation strategy, and updating the warehouse management information; A multi-objective collaborative optimization scheduling unit for comprehensively considering the physical state, environmental influence, historical operation load and life cycle data characteristics of electric power materials, performing risk assessment, predicting the scheduling scene of electric power materials, combining the warehouse management information of each warehouse area, and synchronously achieving the multi-objective collaborative optimization of the shortest path, the shortest total scheduling time and the lowest scheduling energy consumption to pre-generate a scheduling strategy.

21. A computer readable storage medium characterized by, A computer program product for storing computer instructions, which, when executed by a processor, completes the steps in the method of any one of claims 1-20.

22. An electronic device, comprising: A computer program product for storing computer instructions, which, when executed by a processor, completes the steps in the method of any one of claims 1-20.

Images