Closed-loop management method and anomaly early warning system for material work orders in new energy power plants

Abstract

This invention discloses a closed-loop management method and anomaly early warning system for material work orders in new energy power plants. It relates to the field of new energy power plant material management technology and addresses issues such as inaccurate material consumption modeling, distorted merged requisition data, and high false alarm rates in complex scenarios like emergency repairs at mountainous wind farms during winter icing. The method uses historical consumption data to model the average consumption and standard deviation of each material under different work order types, and generates recommended material usage based on a safety factor. It also checks inventory safety margins in real time and provides early warnings of shortage risks. Demands are aggregated within the scheduling cycle, prioritizing internal allocation before external procurement. For scenarios like emergency repairs at mountainous wind farms during winter icing, a repair session mechanism is introduced. Based on multiple factors such as vehicle trajectory, icing intensity, and operation duration, merged requisition records are reconstructed to accurately allocate consumption across work order dimensions and correct usage deviations. A comprehensive anomaly score is calculated using multi-dimensional indicators to achieve tiered early warning.

Description

Technical Field

[0001] This invention relates to the field of new energy power plant material management technology, and more specifically, to a closed-loop management method and anomaly early warning system for new energy power plant material work orders. Background Technology

[0002] Traditional material management systems can generally meet the needs of routine operation and maintenance scenarios, but they expose significant shortcomings under extreme conditions such as winter icing and nighttime emergency repairs at mountainous wind farms. To reduce the number of trips to the warehouse, repair teams often combine materials from multiple work orders and centrally record usage data, leading to a disconnect between requisition records and work orders. Existing systems assume each requisition corresponds to a single work order, failing to handle the structural misalignment of data after combined requisitions. This results in distorted calculations of usage and return deviations, generating numerous false alarms. Furthermore, the system lacks the collection and encoding of contextual information such as vehicle trajectories and work sequences, making it difficult to reconstruct the actual flow path of materials among multiple wind turbines and reducing traceability. Historical consumption statistics are distorted due to data contamination, affecting the accuracy of demand parameters μ and σ, thus weakening the scientific validity of recommended application quantities. In terms of inventory management, existing methods rely on simple book-based verification, with isolated discrepancy analysis that cannot incorporate multi-dimensional risk indicators. Early warning mechanisms are based on fixed thresholds, resulting in a high false alarm rate and a lack of adaptive optimization capabilities, making it difficult to balance repair efficiency and material supply, increasing operation and maintenance costs and safety risks.

[0003] To address the above problems, this invention proposes a solution. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a closed-loop management method and an anomaly early warning system for material work orders of new energy power plants, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a preferred embodiment, it includes:

[0007] Based on historical consumption records, the recommended and final application quantities of materials for work orders are generated. Combined with inventory ledgers and in-transit records, the demand gap is calculated and material shortage warnings, allocation and procurement control are triggered. At the arrival and requisition return stages, quantity deviations are compared to generate arrival acceptance and usage abnormality warnings.

[0008] Emergency repair sessions are introduced based on emergency repair vehicles. The vehicle's driving trajectory is used to obtain the icing intensity, operation time, and road complexity factor. The weights and material allocation ratios are calculated to form the corrected total work order requisition amount. The standardized indicators of usage deviation are combined to distinguish between structural deviations and abnormal usage caused by combined requisition.

[0009] The system calculates anomaly indicators from multiple dimensions, including inventory, demand forecasting, and work order timeliness, and weights them to form a comprehensive anomaly score. It also generates multidimensional anomaly warnings and adjusts the thresholds, weights, and demand forecasting parameters of each indicator.

[0010] In a preferred embodiment, historical consumption records are statistically analyzed by material and work order type to generate recommended application quantity and final application quantity records for work order materials; then, by summarizing inventory and demand data from inventory ledgers, material work orders, parameter configurations, and in-transit records, the total demand for the cycle, the expected remaining inventory, the total amount available for internal allocation, and the gap after internal supply are calculated to generate material shortage early warning candidates, allocation records, and procurement records.

[0011] In a preferred embodiment, the actual quantity received, the planned quantity received, the quantity deviation threshold, and the quality inspection results are compared to determine whether the received materials should be included in the inventory and the corresponding demand record should be marked as supply completed, or the material code, batch number, quantity difference, and quality inspection status should be written into the early warning event database and marked as abnormal arrival acceptance.

[0012] In a preferred embodiment, based on the material work order records that have been marked as completed, material codes and batch information are collected by using barcode or radio frequency identification scanning devices when materials are issued and returned to the warehouse. The actual quantity issued, the actual quantity used, the theoretical quantity to be returned, the actual quantity returned to the warehouse, and the quantity scrapped are recorded in the operation and maintenance management platform, and the inventory ledger is updated and updated simultaneously to achieve batch-based quantity management of work order materials from issuance, use to return and scrapping.

[0013] The system summarizes the total quantity requisitioned, theoretically due quantity to be returned, actual total quantity returned, return deviation, and standardized usage deviation index based on recommended application quantity and historical standard deviation for each batch of records under the work order. It compares the return deviation and usage deviation with the corresponding thresholds, generates abnormal return difference warnings and abnormal usage warnings for work orders and materials that exceed the thresholds, restores inventory for intact returned quantities, deducts inventory from scrapped quantities and writes them to the scrapping record and warning event database.

[0014] In a preferred embodiment, a repair session based on the repair vehicle is introduced. When materials are requisitioned uniformly from the warehouse, records are generated in the repair session record table based on the outbound time, vehicle identification, and work order list, and the total material requisition amount within the session is obtained. Then, using the matching results of vehicle driving trajectory and wind turbine geolocation code, the actual start and end time of each work order in the repair session and the corresponding wind turbine are determined. The icing intensity index is extracted from the operation monitoring system, and combined with the operation duration factor and road complexity factor, a multi-factor set for the winter icing nighttime repair scenario of mountain wind farms is constructed. The three types of factors are normalized and weighted according to the preset weight coefficients in the parameter configuration table to obtain the comprehensive operation weight. Based on this, the material allocation ratio of each work order in the merged requisition is calculated, and the total material requisition amount within the session is allocated to the corrected total requisition amount of the work order to replace the original total requisition amount of the work order.

[0015] In a preferred embodiment, the theoretical return quantity and return deviation are recalculated based on the corrected total work order requisition quantity. This is then compared again using the standardized usage deviation index and the return difference warning threshold to distinguish the statistical structural deviation introduced by merged requisition and centralized backfilling from genuine abnormal usage behavior. Simultaneously, only when the same vehicle uniformly requisitions materials for multiple wind turbines under icing alarm status within a preset short time window, and the material codes and timestamps in the requisition records highly overlap and contain multiple different wind turbine equipment identifiers, is it marked as a repair session and triggers the session-level path reconstruction and allocation correction process for vehicle trajectory matching and multi-factor allocation. Regular single work order material requisition or multi-work order batch material requisition under non-icing conditions continues to use the original processing path of directly generating requisition records and total work order requisition quantity by work order.

[0016] In a preferred embodiment, the inventory discrepancy rate is calculated and an inventory discrepancy warning is generated by comparing the book inventory quantity with the physical inventory quantity in the inventory dimension; in the demand forecasting dimension, a time series forecasting model is constructed based on the actual demand in historical cycles to obtain the cyclical forecasting demand and a demand forecasting deviation warning is generated based on the forecasting error; in the shortage risk dimension, a shortage risk indicator is calculated by using forecasting demand, book inventory, quantity in transit, and safety stock; and in the work order timeliness dimension, the work order delay ratio is calculated by using the planned completion time, actual completion time, and standard duration and a work order delay warning is generated.

[0017] In a preferred embodiment, in the dimension of multidimensional anomaly early warning and closed-loop optimization, a comprehensive anomaly score is formed by weighting the usage deviation standardization index, inventory difference rate, shortage risk index and work order delay ratio. Multidimensional anomaly early warning records are generated according to multi-level thresholds, and the thresholds, weights and demand forecasting parameters of each index are adjusted periodically based on the statistical results of false alarms and missed alarms.

[0018] In a preferred embodiment, it includes: a material demand shortage control module, an emergency repair session path allocation module, a multi-dimensional anomaly early warning closed-loop module, and signal connections between the modules;

[0019] The material demand shortage control module is used to generate the recommended and final application quantities of work order materials based on historical consumption records. It combines inventory ledgers and in-transit records to calculate demand gaps and trigger material shortage warnings, as well as allocation and procurement control. It compares quantity deviations at the arrival and requisition / return stages to generate arrival acceptance and usage abnormality warnings.

[0020] The emergency repair session path allocation module is used to introduce emergency repair sessions on a vehicle-by-vehicle basis. It uses the vehicle's driving trajectory to obtain the icing intensity, operation time and road complexity factor, calculates the weight and material allocation ratio to form the corrected total work order requisition amount, and combines the usage deviation standardization index to distinguish between structural deviations and abnormal usage caused by combined requisition.

[0021] The multidimensional anomaly early warning closed-loop module is used to calculate anomaly indicators in multiple dimensions such as inventory, demand forecasting and work order timeliness, weight them to form a comprehensive anomaly score, generate multidimensional anomaly early warning, and adjust the thresholds, weights and demand forecasting parameters of each indicator.

[0022] The technical effects and advantages of the closed-loop management method and anomaly early warning system for material work orders in new energy power plants of this invention are as follows:

[0023] This invention effectively solves the data distortion problem caused by merged requisition by using emergency repair sessions and multi-factor allocation correction, ensuring accurate consumption at the work order level and improving the reliability of usage and return deviation analysis. Demand modeling, based on historical data and work order types, makes the recommended application quantity more scientific and reduces human error. Multi-dimensional anomaly early warning integrates usage deviation, inventory difference, shortage risk, and work order delay indicators, achieving tiered alarms through weighted scoring, significantly reducing false alarm rates. A closed-loop optimization mechanism dynamically adjusts parameter thresholds and weights based on early warning processing results, enabling the system to adapt to changes in the operating environment and improving robustness. Scheduling decisions combine internal allocation and external procurement to optimize resource allocation and reduce inventory backlog and shortages. Goods arrival acceptance and return management enhance circulation transparency and support precise inventory control. Overall, the system performs exceptionally well in complex scenarios such as mountainous wind farms, improving the accuracy of material management, operational efficiency, and risk response capabilities, while reducing overall operating costs. Attached Figure Description

[0024] Figure 1 This is a sequence diagram of the closed-loop management method and anomaly early warning system for material work orders in new energy power plants according to the present invention.

[0025] Figure 2 This is a schematic diagram of the closed-loop management method and anomaly early warning system module for new energy power plant material work orders of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Example:

[0028] This invention discloses a closed-loop management and anomaly early warning method for material work orders in new energy power plants, such as... Figure 1 As shown, it includes:

[0029] First, based on the records in the historical consumption database, the consumption patterns of different materials under different work order types are modeled. Specifically, using the material code i in the material basic archive database as the index and the work order type c in the historical consumption database as the classification label, the actual consumption quantity field of all records in the historical consumption database that satisfy material code = i and work order type = c is grouped and statistically analyzed. The historical average consumption μ(i,c) and standard deviation σ(i,c) of the material on this type of work order are calculated, and μ(i,c) and σ(i,c) are written into the demand parameter table along with the material code i and the work order type c for storage.

[0030] When on-site maintenance personnel initiate a new maintenance work order, troubleshooting work order, or planned maintenance work order for a specific device through the maintenance management platform, the maintenance management platform assigns a work order number j to the work order and writes the work order type into the work order record. Equipment number and power plant identifier. When maintenance personnel select the required material code i on the work order interface, the maintenance management platform performs the following data processing actions:

[0031] Read the material code i and work order type from the demand parameter table. The corresponding μ(i, ) and σ(i, ), read the corresponding safety factor k(i, ) from the parameter configuration table. The number of recommended applications is calculated using the following formula:

[0032] =μ(i, )+k(i, )·σ(i, );

[0033] in, Let μ(i, ) be the recommended application quantity for material i in work order j; The material type is calculated based on records in the historical consumption database. Historical average consumption; σ(i, ) represents the corresponding historical standard deviation; k(i, The safety factor set for this type of work order is used to allow for fluctuation margins based on average demand.

[0034] The safety factor k(i, The parameters are configured by the operation and maintenance personnel in the parameter configuration table based on historical material shortage rates, work order importance, and on-site experience. If necessary, they can also be adjusted offline based on historical operation data.

[0035] Next, the operations and maintenance personnel checked the number of recommended applications on the interface. Within the floating ratio range defined by the operation and maintenance management platform, you can use the input box to... Make minor adjustments to obtain the final number of applications. The operation and maintenance management platform will use the work order number j and work order type. Equipment number, power plant identification, material code, recommended application quantity and the final number of applications Write a record to the material work order database to generate a structured material work order record.

[0036] During the approval control phase, the operation and maintenance management platform performs the following data reading and comparison operations on each material work order record that is pending approval:

[0037] Read the available inventory of material i in the power plant warehouse during the current scheduling cycle from the inventory ledger database; summarize the occupied quantity of material i for all approved but not yet supplied work orders from the material work order database; and read the safety stock SS(i) corresponding to material i from the parameter configuration table.

[0038] Based on this, the operation and maintenance management platform calculates the current available inventory, the quantity already occupied, and the quantity requested in this work order. The system calculates the expected remaining inventory after approving the work order and compares it with the safety stock SS(i). When the comparison result shows that the expected remaining inventory is less than SS(i), the operation and maintenance management platform displays a material shortage risk warning on the approval interface of the material work order, and writes the material code i, work order number j, and the currently calculated safety stock margin into the warning event database in the form of a warning record, marking it as a material shortage warning candidate.

[0039] Furthermore, the operation and maintenance management platform takes records in the material work order database that are in the "approved and awaiting supply" status as input, and uses a preset scheduling period k as the time unit to calculate the supply and demand. For all work order records within the scheduling period k, the operation and maintenance management platform groups them according to material code i, and calculates the final application quantity for all work orders with the same material code i. Sum the quantities to obtain the total demand R(i,k) for material i within the current period, and temporarily store R(i,k) in memory or a temporary table.

[0040] Subsequently, the operation and maintenance management platform retrieves the available inventory of material i in each power plant warehouse within the enterprise from the inventory ledger database, based on warehouse identifier m and material code i. Simultaneously, it reads the safety stock of materials configured for each warehouse from the parameter configuration table. For each warehouse m and each type of material i, the operation and maintenance management platform calculates the available inventory for material i in that warehouse according to the rule of subtracting safety stock from available inventory. Greater than When, take the difference. - The number of units available for allocation is set as follows; otherwise, the number of units available for allocation is zero.

[0041] Then, the available quantities from each warehouse are summed to obtain the total available quantity of material i within the enterprise during period k. Simultaneously, the quantity of material i that has not yet arrived at the warehouse within the current period is retrieved from the in-transit record database. .

[0042] Based on the above data, the operation and maintenance management platform calculates the supply gap for material i after internal supply, following the order of first using internal inventory and then deciding on external procurement. The specific calculation is as follows:

[0043] ;

[0044] in, R(i,k) represents the remaining gap of material i after internal allocation and in-transit replenishment within the scheduling period k; R(i,k) is the total approved demand for material i within this period. This represents the total amount that can be allocated within the company during the current period. This represents the quantity of supplies en route for the current period. The operations and maintenance management platform uses this information... The result of implementing supply decisions: when When it is zero, it means that the demand can be met solely through internal allocation and in-transit replenishment; when A value greater than zero indicates that there is still a shortage of materials that need to be met through external procurement. It is directly used as the quantity for external procurement.

[0045] After that, regarding the passage To cover a portion of the needs, the operation and maintenance management platform, based on the available quantity of each warehouse and its geographical distance from the target power plant or allocation priority, [determines / adjusts / adjusts / etc.]. The allocation process is as follows: The available quantities in each warehouse are sorted from largest to smallest, and then allocated to the total periodic demand R(i,k) in descending order until R(i,k) is completely covered or the available quantities are exhausted. Each allocation operation generates an allocation record, which includes fields such as the sending warehouse identifier, receiving power station identifier, material code i, allocation quantity, and planned shipment time. The allocation record is written to the material work order database, and the allocated quantity for the corresponding warehouse is deducted from the available inventory in the inventory ledger database, while simultaneously being marked with a pending shipment status.

[0046] for For materials with a value greater than zero, the operation and maintenance management platform will... As the purchase quantity, a purchase record is generated. The purchase record includes fields such as material code i, purchase quantity, required delivery time, and supplier identifier. It is written to the purchase record database and linked with the relevant material work order record.

[0047] When materials prepared through allocation or procurement actually arrive at the warehouse, warehouse management personnel use barcode or RFID scanning equipment to read the material code i and batch number b for each batch of materials at the arrival site, and enter the actual quantity of materials received through the arrival acceptance interface. The operations and maintenance management platform reads the planned delivery quantity for that batch from the material work order database or procurement record database. Calculate the quantity difference - The difference is then compared with the quantity deviation threshold configured for that material in the parameter configuration table to determine if there is a quantity anomaly in the batch. Simultaneously, quality inspectors conduct performance testing and visual inspection of the batch of materials according to the company's established acceptance standards. The test results are entered into the acceptance interface as qualified / unqualified status fields, written into the acceptance record, and associated with batch number b.

[0048] For batches where the quantity difference is within the threshold and the quality inspection status is qualified, the operation and maintenance management platform will record the actual quantity received. The batch of materials is recorded in the inventory ledger database and added to the available inventory. The corresponding demand record in the material work order database is marked as supplied completed. For batches with quantity discrepancies exceeding the threshold or with a quality inspection status of unqualified, the operation and maintenance management platform writes the material code i, batch number b, quantity discrepancy, and quality inspection status to the early warning event database, marks it as an abnormal arrival and acceptance, and does not add the available inventory of this batch of materials to the inventory ledger database.

[0049] Furthermore, based on the aforementioned material work order records that have been marked as completed, when a certain work order j enters the execution phase, the on-site maintenance personnel submit a material requisition request in the maintenance management platform. The warehouse management personnel determine the available material code i and the maximum quantity based on the material work order record associated with that work order, and use barcode or RFID scanning equipment on-site to read the material code i and batch number b of each outbound material, and enter the actual quantity requisitioned through the outbound interface. The operations and maintenance management platform, while writing a requisition record to the material work order database, simultaneously subtracts the available inventory quantity of the corresponding batch b from the inventory ledger database. This reflects the actual flow of materials from the warehouse to the site.

[0050] After the work order is completed, the on-site maintenance personnel fill in the actual quantity of each material used in the work order completion interface. This field, along with the work order number j and material code i, is written to the material work order database. The operation and maintenance management platform aggregates all batches of requisition records under work order j by material code i, and identifies the requisition records corresponding to the same material i. Sum the quantities to get the total quantity of material i requisitioned for this work order. Subsequently, the operation and maintenance management platform will determine the total number of units used. and actual usage quantity Calculate the theoretical amount to be refunded:

[0051] = - ;

[0052] in, This indicates the quantity of material i that work order j should theoretically be returned to the warehouse. The operations and maintenance management platform will... Write the information into the material work order database and generate a list of materials to be returned for reference when returning materials to the warehouse.

[0053] Next, the maintenance personnel returned the unused and intact materials to the warehouse according to the list of materials to be returned. During the return process, warehouse management personnel also used a scanning device to read the material code "i" and batch number "b" of the returned materials, and entered the actual quantity returned through the return interface. The operations and maintenance management platform will record the actual number of returned goods. Write the data to the material work order database and increase the available inventory quantity for the corresponding batch in the inventory ledger database. For multiple return records generated by the same work order j at different times or in different batches b, the operation and maintenance management platform will, at the work order level, record all return records under that work order. Sum the results and treat the total value as the actual total amount of material i returned to the warehouse by work order j.

[0054] Subsequently, the operation and maintenance management platform calculated the theoretically required refund amount. Compared with the aforementioned actual total amount of goods returned The difference between the two values ​​is used to calculate the return deviation, which is then compared with the threshold for return deviation in the parameter configuration table. When the deviation exceeds the threshold, the operation and maintenance management platform writes the material code i, work order number j, and deviation value into the early warning event database, marking it as an abnormal return difference.

[0055] For materials that are damaged on-site or deemed unusable upon inspection, warehouse management personnel mark them as scrap materials in the return interface and enter the quantities returned in good condition and the quantities scrapped separately. The operation and maintenance management platform only restores inventory for the quantities in good condition, and deducts inventory from the inventory ledger database for the scrapped quantities, writing the scrapped quantities and corresponding material codes i into the scrapped record database;

[0056] In terms of anomaly analysis of material usage, the operation and maintenance management platform uses the actual usage quantity in the work order completion record. The number of recommended applications calculated in step one Based on this, the usage deviation is calculated, and combined with the historical consumption database of material i in the work order type. The standard deviation σ(i, Normalization is performed to obtain the standardized index of usage deviation. Preferably, it can be defined as:

[0057] ;

[0058] Where ε is a very small positive number stored in the parameter configuration table, when | When the usage deviation exceeds the pre-set threshold in the parameter configuration table, the operation and maintenance management platform records a usage anomaly warning in the warning event database. Work orders and material combinations that show usage anomalies multiple times can be marked as objects with abnormal usage behavior.

[0059] It should be noted that in nighttime repair scenarios at mountainous wind farms after winter icing, maintenance teams often use a single pickup truck or tracked vehicle to carry shared supplies such as safety ropes, heating blankets, de-icing agent sprayers, and temporary lighting equipment to multiple wind turbines for blade de-icing and tower external structural reinforcement under conditions of strong winds, low temperatures, and icy roads. To reduce the number of trips to the warehouse and avoid frequent loading and unloading on icy roads, the teams often collect all the materials needed for multiple work orders from the warehouse at once before departure, generating only one consolidated requisition record on the operation and maintenance management platform. However, during actual on-site operations, each wind turbine corresponds to an independent maintenance work order, and the specific amount of materials used is temporarily recorded by the team leader in a paper logbook or mobile phone memo. After the repair is completed and the network is restored to stability, the materials are then entered into the operation and maintenance management platform one by one according to the work orders. Work order completion information.

[0060] In the above scenario, since the merged requisition records only correspond to one or two master work orders, the remaining slave work orders only have usage quantities recorded in the system. However, there is a lack of association with batch number b. The details, coupled with unstable network conditions in mountainous areas at night leading to misalignments between uploaded timestamps and actual work sequences, resulted in a data structure in the material work order database characterized by ambiguous requisition paths, delayed usage updates, and broken work order-batch relationships. Essentially, this problem stems from an inconsistency between the data collection method at the perception layer and the merged operation mode at the execution layer. , , The statistical representation of "equal quantity" does not indicate the actual flow path of materials, but rather a mixed structure of materials being consolidated and requisitioned, then split and backfilled according to work orders.

[0061] In existing technologies, material management systems typically assume that each requisition record corresponds to a single work order, and force users to select work order number j on the requisition interface, and directly generate the requisition record based on the work order. For instances of combined requisitioning and centralized data entry at emergency repair sites, the system primarily addresses these issues by manually rewriting work order numbers, splitting inbound and outbound slips, or manually adjusting inventory. It fails to provide structured modeling for the unique pattern of vehicle-level combined requisitioning—multiple work orders used sequentially—and centralized data entry across time periods. On one hand, based on… , , as well as Calculate usage deviation and return deviation, and generate accordingly. When issuing warnings for discrepancies in returns, it is impossible to distinguish between genuine abnormal usage and record misalignments caused by combined requisitions, easily leading to a large number of false usage deviations or abnormal return discrepancy alerts. On the other hand, when subsequently summarizing material consumption by work order dimension and calculating historical consumption averages and fluctuations, the master work order for combined requisitions is aggregating excessive requisition and usage behaviors, while the slave work orders only retain the data filled in afterward. This results in a significant discrepancy between the material consumption distribution across work orders and the actual situation on site, thereby affecting the accuracy of μ(i,c), σ(i,c), and related usage deviation judgments based on historical consumption statistics.

[0062] Furthermore, because the existing system lacks the collection and encoding of contextual information such as vehicle travel trajectory, operating fan sequence, and offline cache upload batches in the requisition logs, it is difficult to reconstruct the actual path of materials flowing between multiple fan bases from the data level during post-event audits or anomaly analysis. Even when combined with... , , Static records are insufficient to reconstruct the actual material usage chain for each work order and each wind turbine on the night of the emergency repair, resulting in a significant decrease in the explainability and traceability of closed-loop material management in such extreme scenarios.

[0063] Therefore, in this embodiment, based on the aforementioned usage deviation analysis, for material work orders identified as emergency nighttime repairs for winter icing at mountainous wind farms and involving combined requisitioning, a repair session S based on repair vehicles is introduced. Combining multi-factor information such as vehicle travel trajectories, wind turbine operation sequences, and on-site icing intensity, session-level path reconstruction and allocation correction are performed on the combined requisition records to restore... , Authenticity at the work order level. Specifically, in the warehouse outbound process, when a team uses the same vehicle to collect materials for multiple ice-covered emergency repair work orders within a short period of time, the operation and maintenance management platform generates a merged requisition record and, based on the outbound time, vehicle identifier, and work order list, generates an emergency repair session S. It then writes S, along with the material code i, the quantity requisitioned, and the vehicle identifier involved in this requisition, into the emergency repair session record table. After the emergency repair vehicle leaves the warehouse, the positioning terminal installed on the vehicle continuously collects location point sequences, uploading or temporarily storing the timestamps and latitude / longitude trajectories to form a vehicle driving trajectory sequence. This trajectory sequence is matched with the wind turbine's geographical location code through a geographical matching relationship, ensuring that each time a vehicle stays near a wind turbine, the time interval can be marked as a work segment corresponding to a specific work order j within the emergency repair session S.

[0064] After the emergency repairs were completed and the network was restored, the team leader entered the missing data item by item according to the work order. Based on the completion time, the operation and maintenance management platform uses the emergency repair session S as the key to read the total amount of materials requisitioned for this merged requisition from the emergency repair session record table. The system reads the completion time and equipment identifier of each work order j in the emergency repair session S from the work order record, and uses the matching results of the vehicle travel trajectory and the wind turbine location to determine the actual start and end time of each work order j in the emergency repair session S. , Create a record of the corresponding operation segment with the ventilation unit.

[0065] Furthermore, to reflect the potential differences in material consumption among different wind turbines during the same emergency repair session S, the operation and maintenance management platform extracts the material consumption data for each wind turbine from the wind farm operation monitoring system. to Ice intensity index within a time period This indicator can be calculated based on blade vibration characteristics, current power deviation, or blade surface temperature sensor data; simultaneously, the duration Δt(j,S) of each work segment within the emergency repair session S is used as the metric. - The system represents the operation duration factor and pre-configures a road complexity factor for each wind turbine based on the slope of the terrain, the level of road icing and snow, or the difficulty of nighttime passage. The above three types of quantities together constitute a multi-factor set for the nighttime emergency repair scenario of a mountainous wind farm after icing in winter, which is unique to this embodiment.

[0066] In obtaining Δt(j,S) of all work orders j within the emergency repair session S, and Subsequently, the operation and maintenance management platform normalizes the above factors. Preferably, the operation time, icing intensity, and road complexity can be linearly scaled based on the minimum and maximum values ​​of each factor within the emergency repair session S. The operation time factor, icing intensity factor, and road complexity factor of each work order within the emergency repair session S are scaled to the [0,1] interval, and the weighting coefficients preset in the parameter configuration table for the nighttime emergency repair scenario after winter icing in mountainous wind farms are applied. Calculate the overall task weight:

[0067] ;

[0068] Subsequently, using the sum of β(j,S) of all work orders within the emergency repair session S as the normalization factor, the material allocation ratio of each work order in this consolidated requisition was calculated. Based on this ratio, the total usage of session-level data will be... Break down the work order by work order dimension to generate the corrected work order usage quantity:

[0069] ;

[0070] Then Aggregate material code i to work order j dimension to obtain This is used to replace the original method of directly summing up the requisition records of some master work orders. .

[0071] In obtaining Afterwards, the operation and maintenance management platform maintains the actual number of users manually added to the work order completion record. Remain unchanged, recalculate the revised theoretical refund amount for the session scenario. and the actual number of returned goods The comparison yields the corrected return deviation index. For work order combinations identified as nighttime emergency repair scenarios following winter icing at mountain wind farms and involving combined requisitioning, the deviation standardization index is used. And the judgment of the return difference warning is no longer directly based on the original , , but based on , By recalculating and comparing with the threshold, the statistical structural bias introduced by merged requisition and centralized backfilling can be distinguished from genuine abnormal usage behavior at the indicator level within the operation and maintenance management platform.

[0072] Meanwhile, to ensure that the above-mentioned session-level reconstruction does not introduce additional complexity in ordinary non-merged material requisition scenarios, this embodiment introduces multi-factor conditions when identifying emergency repair session S: it is only marked as emergency repair session S and the vehicle trajectory matching and multi-factor allocation process is triggered when the following conditions are met simultaneously: the same vehicle uniformly requisitions materials for the work orders corresponding to multiple wind turbines in the icing alarm state within a preset short time window; the material codes and timestamps in the material requisition records highly overlap; and the work order list contains multiple different wind turbine equipment identifiers. For regular single work order material requisition or multi-work order batch material requisition under non-icing conditions, the original method of directly generating materials by work order continues to be used. The processing path will not initiate the aforementioned correction algorithm.

[0073] Furthermore, in terms of inventory, the operations and maintenance management platform reads the book inventory quantity from the inventory ledger database by material code i. Read the physical quantity of items collected during on-site inventory counting in the same scheduling cycle k from the inventory record database. Inventory of physical quantities Originating from the inventory count plan: During the inventory count, warehouse management personnel use inventory terminals to scan the location and material codes one by one, enter the physical quantity, and upload it to the inventory record database. The operations and maintenance management platform then... and Calculate the inventory discrepancy rate:

[0074] ;

[0075] in, This defines the inventory discrepancy rate for material i during scheduling period k. The parameter configuration table sets an inventory discrepancy threshold for each material; when | When the threshold is exceeded, the operation and maintenance management platform will write the material code i and the corresponding warehouse identifier into the early warning event database and generate an inventory difference early warning record.

[0076] In terms of demand forecasting, the operations and maintenance management platform uses the actual demand D(i,k) statistically analyzed periodically from the historical consumption database as a basis, and employs time series methods, such as moving averages or exponential smoothing, to forecast the demand for each type of material. Taking exponential smoothing as an example, the forecast value from the previous period can be used... Weigh the forecasted demand for the next period by the actual demand D(i,k) of the current period. and will Write the forecast record into the demand forecast database as the forecast demand for period k+1.

[0077] Subsequently, after period k ends, the demand forecast records stored in the database are used for... The system takes the actual demand D(i,k) in the historical consumption database as input, calculates the prediction error for period k, and statistically analyzes the average error and error fluctuation over multiple periods. When it is found that the prediction error of a certain material deviates from zero for a long time and exceeds the tolerance range set in the parameter configuration table, the operation and maintenance management platform records a demand prediction deviation warning in the early warning event database and prompts that the prediction parameters of the material need to be adjusted.

[0078] In terms of shortage risk analysis, the operations and maintenance management platform combines the predicted demand in the demand forecasting database. Book inventory in the inventory ledger database The number of items in transit in the in-transit record database And the safety stock SS(i) in the parameter configuration table, calculate the shortage risk index of material i in period k according to the following relationship. :

[0079] ;

[0080] in, This reflects whether there is a gap between current inventory and the quantity in transit, and its relative size, taking into account forecasted demand and safety stock.

[0081] In terms of work order timeliness, the operation and maintenance management platform uses the planned completion time of work orders in the material work order database or work order management records. Actual completion time of work order And the parameter configuration table specifies the corresponding work order type. Pre-set standard construction period Input the work order delay rate. , can be defined as:

[0082] ;

[0083] in, Used to measure the delay level of work order j, when When the time limit is exceeded, the work order is considered an abnormal time-delay work order and can be used as a basis for work order delay warning.

[0084] In terms of generating multi-dimensional anomaly warnings, the operation and maintenance management platform will use standardized deviation indicators for each type of material i and each work order j. Inventory discrepancy rate Shortage risk indicators and the percentage of work orders delayed The features are combined into a feature vector, and a weighted summation method is used to calculate the comprehensive anomaly score:

[0085] ;

[0086] Wherein, Score(i,k,j) is the comprehensive anomaly score of material i in the dimensions of scheduling period k and work order j; w1, w2, w3, w4 are the weights of each indicator stored in the parameter configuration table, which are used to reflect the degree of contribution of different anomaly types to the overall risk.

[0087] The weights of the above indicators and the multi-level early warning thresholds corresponding to Score(i,k,j) can be set by operation and maintenance personnel based on the false alarm rate, missed alarm situation and on-site operation and maintenance experience in the historical early warning records. They can also be adjusted regularly in combination with closed-loop statistical results during system operation to adapt to the changes in the risk characteristics of material management in new energy power plants.

[0088] Next, the operation and maintenance management platform classifies Score(i,k,j) according to the preset multi-level thresholds, divides the score into different warning levels, and writes the warning level, main indicator values, material code i, and work order number j into the warning event database to form a multi-dimensional abnormal warning record.

[0089] Finally, during the closed-loop optimization process, the operation and maintenance management platform regularly performs statistical analysis on the early warning records and processing results in the early warning event database. When a certain type of early warning occurs frequently and is found to be mostly false alarms after manual verification, the corresponding thresholds and weights w1, w2, w3, and w4 in the parameter configuration table are adjusted according to the cause category registered in the processing records. For example, the deviation threshold may be appropriately relaxed or the weight of this type of indicator may be reduced. When a certain type of risk has a serious impact on actual operation but the early warning trigger is not sensitive, the threshold of this type of indicator is tightened or the weight is increased accordingly. For materials with large demand forecast deviations, μ(i,c) and σ(i,c) are re-estimated or the smoothing parameters in the time series forecast are adjusted based on the statistical results of the forecast error, so that the parameters used in the above-mentioned material demand modeling can reflect the latest operational data. For warehouses and materials with consistently high inventory difference rates and return deviations, the inventory frequency or material management rules can be adjusted accordingly.

[0090] This invention also proposes a closed-loop management and anomaly early warning system for material work orders in new energy power plants, such as... Figure 2 As shown, it includes: a material demand shortage control module, an emergency repair session path allocation module, a multi-dimensional anomaly early warning closed-loop module, and signal connections between the modules;

[0091] The material demand shortage control module is used to generate the recommended and final application quantities of work order materials based on historical consumption records. It combines inventory ledgers and in-transit records to calculate demand gaps and trigger material shortage warnings, as well as allocation and procurement control. It compares quantity deviations at the arrival and requisition / return stages to generate arrival acceptance and usage abnormality warnings.

[0092] The emergency repair session path allocation module is used to introduce emergency repair sessions on a vehicle-by-vehicle basis. It uses the vehicle's driving trajectory to obtain the icing intensity, operation time and road complexity factor, calculates the weight and material allocation ratio to form the corrected total work order requisition amount, and combines the usage deviation standardization index to distinguish between structural deviations and abnormal usage caused by combined requisition.

[0093] The multidimensional anomaly early warning closed-loop module is used to calculate anomaly indicators in multiple dimensions such as inventory, demand forecasting, and work order timeliness, weight them to form a comprehensive anomaly score, generate multidimensional anomaly early warning, and adjust the thresholds, weights, and demand forecasting parameters of each indicator.

[0094] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0095] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0096] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0097] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0098] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0099] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A closed-loop management and anomaly early warning method for material work orders in new energy power plants, characterized in that, include: Based on historical consumption records, the recommended and final application quantities of materials for work orders are generated. Combined with inventory ledgers and in-transit records, the demand gap is calculated and material shortage warnings, allocation and procurement control are triggered. At the arrival and requisition return stages, quantity deviations are compared to generate arrival acceptance and usage abnormality warnings. Emergency repair sessions are introduced on a vehicle-by-vehicle basis. These sessions are generated based on outbound time, vehicle identification, and work order list. The vehicle's driving trajectory is used to obtain icing intensity, operation time, and road complexity factors. Weights and material allocation ratios are calculated to form the corrected total work order requisition amount. The usage deviation standardization index is combined to distinguish between structural deviations such as record misalignment caused by merged requisition and abnormal usage. The system performs multi-dimensional calculations on inventory, demand forecasting, and work order timeliness, including the use of deviation standardization indicators, inventory difference rate, shortage risk indicators, and work order delay ratios. These abnormal indicators are weighted to form a comprehensive abnormal score, and multi-dimensional abnormal warnings are generated. The system also adjusts the thresholds, weights, and demand forecasting parameters of each indicator.

2. The closed-loop management and anomaly early warning method for new energy power plant material work orders according to claim 1, characterized in that: The historical consumption records are statistically analyzed by material and work order type to generate records of recommended application quantities and final application quantities for work order materials. Then, by summarizing the inventory and demand data in the inventory ledger, material work orders, parameter configurations, and in-transit records, the total demand for the cycle, the estimated remaining inventory, the total amount that can be allocated within the enterprise, and the gap after internal supply are calculated, and material shortage early warning candidates, allocation records, and procurement records are generated.

3. The closed-loop management and anomaly early warning method for material work orders in new energy power plants according to claim 2, characterized in that: By comparing the actual quantity received, the planned quantity received, the quantity deviation threshold, and the quality inspection results, it is determined whether to include the received materials in the inventory and mark the corresponding demand record as supply completed, or to write the material code, batch number, quantity difference, and quality inspection status into the early warning event database and mark it as an abnormal delivery acceptance.

4. The closed-loop management and anomaly early warning method for new energy power plant material work orders according to claim 3, characterized in that: Based on the material work order records that have been marked as completed, material codes and batch information are collected by using barcode or radio frequency identification scanning devices when materials are issued and returned to the warehouse. The actual quantity issued, actual quantity used, theoretical quantity to be returned, actual quantity returned to the warehouse, and scrapped quantity are recorded in the operation and maintenance management platform, and the inventory ledger is updated and updated in sync. This enables batch-based quantity management of work order materials from issuance and use to return and scrapping. The system summarizes the total quantity requisitioned, theoretically due quantity to be returned, actual total quantity returned, return deviation, and standardized usage deviation index based on recommended application quantity and historical standard deviation for each batch of records under the work order. It compares the return deviation and usage deviation with the corresponding thresholds, generates abnormal return difference warnings and abnormal usage warnings for work orders and materials that exceed the thresholds, restores inventory for intact returned quantities, deducts inventory from scrapped quantities and writes them to the scrapping record and warning event database.

5. The closed-loop management and anomaly early warning method for new energy power plant material work orders according to claim 4, characterized in that: A repair session based on the vehicle is introduced. When materials are requisitioned from the warehouse, records are generated in the repair session record table based on the outbound time, vehicle identification, and work order list, and the total material requisition quantity within the session is obtained. Then, by matching the vehicle's driving trajectory with the wind turbine's geographical location code, the actual start and end times of each work order in the repair session and the corresponding wind turbine are determined. The icing intensity index is extracted from the operation monitoring system, and combined with the operation duration factor and road complexity factor, a multi-factor set is constructed for the winter icing nighttime repair scenario of mountain wind farms. The three types of factors are normalized and weighted according to the preset weight coefficients in the parameter configuration table to obtain the comprehensive operation weight. Based on this, the material allocation ratio of each work order in the merged requisition is calculated, and the total material requisition quantity within the session is allocated to the corrected total requisition quantity of the work order to replace the original total requisition quantity of the work order.

6. The closed-loop management and anomaly early warning method for new energy power plant material work orders according to claim 5, characterized in that: Based on the revised total work order requisition amount, the theoretical return quantity and return deviation are recalculated, and compared again with the standardized indicators of usage deviation and the threshold of return difference warning. This distinguishes the statistical structural deviation introduced by merged requisition and centralized backfilling from the real abnormal usage behavior. At the same time, only when the same vehicle requisitions materials for multiple wind turbines in the icing alarm state within a preset short time window, and the material code and timestamp in the material requisition record highly overlap and contain multiple different wind turbine equipment identifiers, is it marked as a repair session and triggers the session-level path reconstruction and allocation correction process of vehicle trajectory matching and multi-factor allocation. Regular single work order material requisition or multi-work order batch material requisition under non-icing conditions continues to use the original processing path of directly generating requisition records and total requisition amount of work orders.

7. The closed-loop management and anomaly early warning method for new energy power plant material work orders according to claim 6, characterized in that: In terms of inventory, the inventory discrepancy rate is calculated by comparing the book inventory quantity with the physical inventory quantity, and an inventory discrepancy warning is generated. In the demand forecasting dimension, a time series forecasting model is constructed based on the actual demand in historical cycles to obtain the cyclical forecasting demand and generate a demand forecasting deviation warning based on the forecasting error; in the shortage risk dimension, a shortage risk indicator is calculated by using forecasting demand, book inventory, quantity in transit and safety stock. In terms of work order timeliness, the work order delay ratio is calculated based on the planned completion time, actual completion time, and standard duration, and a work order delay warning is generated.

8. The closed-loop management and anomaly early warning method for new energy power plant material work orders according to claim 7, characterized in that: In the multidimensional anomaly early warning and closed-loop optimization dimension, a comprehensive anomaly score is formed by weighting the usage deviation standardization index, inventory difference rate, shortage risk index and work order delay ratio. Multidimensional anomaly early warning records are generated according to multi-level thresholds. The thresholds, weights and demand forecasting parameters of the usage deviation standardization index, inventory difference rate, shortage risk index and work order delay ratio, as well as the statistical results of false alarms and missed alarms, are adjusted periodically based on the statistical results of false alarms and missed alarms.

9. A closed-loop management and anomaly early warning system for material work orders in new energy power plants, characterized in that: include: The module includes a material demand shortage control module, an emergency repair session path allocation module, a multi-dimensional anomaly early warning closed-loop module, and signal connections between the modules. The material demand shortage control module is used to generate the recommended and final application quantities of work order materials based on historical consumption records. It combines inventory ledgers and in-transit records to calculate demand gaps and trigger material shortage warnings, as well as allocation and procurement control. It compares quantity deviations at the arrival and requisition / return stages to generate arrival acceptance and usage abnormality warnings. The emergency repair session path allocation module is used to introduce emergency repair sessions based on emergency repair vehicles. The emergency repair session is generated based on the outbound time, vehicle identification and work order list. The icing intensity, operation time and road complexity factor are obtained by using the vehicle driving trajectory. The weight and material allocation ratio are calculated to form the corrected total work order requisition. The module also combines the usage deviation standardization index to distinguish between the structural deviation of record misalignment caused by merged requisition and abnormal usage. The multidimensional anomaly early warning closed-loop module is used to perform multidimensional calculations on inventory, demand forecasting, and work order timeliness, including using deviation standardization indicators, inventory difference rate, shortage risk indicators, and work order delay ratio, to form a weighted comprehensive anomaly score, generate multidimensional anomaly early warnings, and adjust the thresholds, weights, and demand forecasting parameters of each indicator.

Images