Data processing method and device, computer equipment and readable storage medium

Abstract

The invention is suitable for the technical field of computers, and provides a data processing method and device, and the method comprises the steps: obtaining a plurality of pieces of initial receiving data matched with a target spare part code in a specified statistical time period; a material receiving mode of the initial receiving data is obtained, data integration processing matched with the material receiving mode is carried out on the initial receiving data, integrated receiving data is obtained, and the receiving quantity of the integrated receiving data is a non-negative number; performing type identification matched with the material picking mode on the integrated picking data to obtain an identification result; and if the identification result represents that the integrated receiving data is interference receiving data, removing the interference receiving data to obtain target receiving data. According to the invention, the efficiency and accuracy of spare part receiving data analysis can be improved.

Description

Technical Field

[0001] The present application belongs to the field of computer technology, and in particular, relates to a data processing method, device, electronic device and readable storage medium. Background Art

[0002] Nuclear power spare parts in nuclear power plants have the characteristics of large number of varieties, small spare parts consumption, long spare parts procurement cycle, and difficult to predict the demand for spare parts. Spare parts are usually collected according to work orders and material collection orders, and may be collected multiple times at the same time. Due to the characteristics of nuclear power spare parts, it is usually necessary to summarize the historical collection data of spare parts by year or month for inventory strategy model to calculate inventory parameters.

[0003] Since spare parts may be returned, and the time between picking and returning may span months or years, when the number of returned parts is greater than the picking data in a certain statistical period, the number of spare parts picked will be negative, and the data needs to be modified (otherwise the inventory strategy model cannot be calculated). When the picking data in a certain statistical period is currently monitored to be negative, the value will be forcibly assigned to 0, that is, there is no picking data in this statistical period.

[0004] Since nuclear power plants will carry out maintenance activities such as transformation, replacement, and defect elimination, the probability of subsequent use of spare parts in the same scale is extremely low for spare parts used in such maintenance activities, and such "special" use data needs to be eliminated. Currently, the use data of such work orders is mainly identified and eliminated based on manual experience. Limited by user experience, the effectiveness of data elimination is inconsistent, and the overall efficiency is low, which affects the promotion and use of nuclear power spare parts inventory strategy model tools. Summary of the invention

[0005] In view of this, the embodiments of the present application provide a data processing method, device, electronic device and readable storage medium, which can improve the efficiency and accuracy of spare parts requisition data analysis.

[0006] A first aspect of an embodiment of the present application provides a data processing method, including:

[0007] Obtain multiple pieces of initial requisition data matching the target spare part code within the specified statistical time period;

[0008] Acquire the material collection method of the initial collection data, and perform data integration processing on the initial collection data adapted to the material collection method to obtain integrated collection data, wherein the collection quantity of the integrated collection data is a non-negative number;

[0009] Performing type identification on the integrated requisition data that is compatible with the requisition method to obtain an identification result;

[0010] If the identification result indicates that the integrated collection data is interference collection data, the interference collection data is removed to obtain target collection data.

[0011] In an implementation of the first aspect, the material picking method is work order material picking, and the data integration processing adapted to the material picking method is performed on the initial picking data to obtain the integrated picking data, including:

[0012] Obtain at least one maintenance work order number attached to the initial collection data;

[0013] Determine a first collection data group corresponding to each maintenance work order number from the plurality of initial collection data;

[0014] Data integration is performed on each of the first collection data groups to obtain integrated collection data corresponding to each of the maintenance work order numbers; wherein the integrated collection data includes a target collection quantity and a target collection time for the spare parts, the target collection quantity is the difference between the total collection quantity for the spare parts in the first collection data group and the total return quantity for the spare parts, and the collection time is the first collection time for the spare parts in the maintenance work order.

[0015] In an implementation of the first aspect, the material picking method includes non-work order material picking, and the data integration processing adapted to the material picking method on the initial picking data to obtain the integrated picking data includes:

[0016] Obtaining at least one batch number associated with the initial requisition data, wherein the batch number is set when the spare parts are accepted for storage;

[0017] Determine a second collection data group corresponding to each batch number from the plurality of initial collection data;

[0018] For each second requisition data group, it is sorted in reverse order according to the material return time. Starting from the first requisition data after sorting, the subsequent requisition data are filled with returned materials in turn until the requisition quantity for spare parts is a non-negative number, and an integrated second requisition data group is obtained, and each integrated second requisition data group is determined as the integrated requisition data.

[0019] In an implementation of the first aspect, the material collection method is work order material collection, and the integrated collection data is linked to the maintenance work order number.

[0020] The step of performing type identification on the integrated requisition data that is compatible with the requisition method to obtain an identification result includes:

[0021] Acquire a first keyword and a second keyword for identifying the interference data, wherein the first keyword has a higher accuracy rate in identifying the interference data than the second keyword;

[0022] Obtaining description keywords of the maintenance work order to which the integrated requisition data is attached;

[0023] Determine an association relationship between the description keyword, the first keyword, and the second keyword;

[0024] Based on the association relationship, the integrated collection data is identified by type to obtain an identification result.

[0025] In an implementation of the first aspect, the performing type identification on the integrated requisition data based on the association relationship to obtain an identification result includes:

[0026] If the association relationship indicates that the description keyword includes the first keyword, then an identification result is obtained that the integrated collection data is interference collection data;

[0027] If the association relationship represents that the description keyword does not include the first keyword but includes the second keyword, then type identification is performed on the integrated requisition data through a type identification model to obtain an identification result, and the identification result indicates whether the integrated requisition data is interference requisition data;

[0028] If the association relationship represents that the description keyword does not include the first keyword and does not include the second keyword, then an identification result is obtained that the integrated collection data is non-interference collection data.

[0029] In an implementation of the first aspect, the type recognition model is a binary classification model, the type recognition model is trained based on the collection data sample, the label of the collection data sample represents whether it is interference collection data, and the sample feature of the collection data sample is obtained by normalizing the collection data.

[0030] In an implementation of the first aspect, the method further includes: acquiring an attribute type of each attribute of the requisition data;

[0031] If the attribute type is a discrete attribute, the discrete attribute value is normalized based on a preset normalization condition to obtain a target attribute value;

[0032] If the attribute type is a numerical attribute, a linear transformation method is used to normalize the attribute value to obtain a target attribute value.

[0033] In an implementation of the first aspect, the acquiring of a plurality of pieces of initial requisition data matching the target spare part code within a specified statistical time period includes:

[0034] Based on the target spare part code, the merged first code and the second code before replacement are traced back step by step in the order of material collection time to obtain a spare part code set, wherein the spare part code set includes the target spare part code, the first code and the second code;

[0035] Based on the spare part code set, a plurality of initial requisition data matching the target spare part code within a specified statistical time period are obtained.

[0036] A second aspect of an embodiment of the present application provides a data processing device, including:

[0037] An acquisition module is used to acquire multiple pieces of initial requisition data matching the target spare part code within a specified statistical time period;

[0038] A data integration module is used to obtain the material collection method of the initial collection data, and perform data integration processing on the initial collection data adapted to the material collection method to obtain integrated collection data, wherein the collection quantity of the integrated collection data is a non-negative number;

[0039] An identification module, used for identifying the type of the integrated requisition data that is compatible with the requisition method, and obtaining an identification result;

[0040] The removal module is used to remove the interfering collection data to obtain the target collection data if the recognition result indicates that the integrated collection data is interfering collection data.

[0041] A fifth aspect of an embodiment of the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the first aspect when executing the computer program.

[0042] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0043] In the first aspect of the embodiment of the present application, by focusing on the collection data that matches the specific target spare parts code within the specified statistical time period, the pertinence and accuracy of the analysis data are ensured, and the interference of irrelevant data is avoided. The data is adaptively integrated and processed according to different collection methods, so that the data format is unified, which is convenient for subsequent analysis and processing, and the efficiency and accuracy of data processing are improved. Through type recognition technology, it is possible to automatically identify which collection data may be interference data, which can significantly improve data quality. After identifying and removing interfering collection data, the target collection data set obtained is purer, which can effectively improve the efficiency and accuracy of spare parts collection data analysis.

[0044] It can be understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 is a schematic diagram of a system structure for implementing a data processing method provided in an embodiment of the present application;

[0047] Figure 2 It is a flow chart of a method for determining nuclear power spare parts inventory parameters provided in an embodiment of the present application;

[0048] Figure 3 It is a schematic diagram of the implementation flow of the data processing method provided in the embodiment of the present application;

[0049] Figure 4 It is a flowchart of a data integration method for work order collection provided in an embodiment of the present application;

[0050] Figure 5 It is a flowchart of a data integration method for non-work order collection provided in an embodiment of the present application;

[0051] Figure 6 is a flow chart of a method for identifying interference with data collection based on keywords provided in an embodiment of the present application;

[0052] Figure 7 It is a bubble chart of interference collection data analysis based on keyword information provided in an embodiment of the present application;

[0053] Figure 8 This is a distribution diagram of the accuracy of identifying special types of use based on keywords provided in an embodiment of the present application;

[0054] Fig. 9 It is a flow chart of a method for identifying the use data based on the association relationship provided in an embodiment of the present application;

[0055] Fig.10 is a flow chart of a genetic algorithm provided in an embodiment of the present application;

[0056] Fig.11 It is a normalized operation flow chart of the collection data provided in the embodiment of the present application;

[0057] Fig.12This is an example diagram of the relationship between the spare parts collection time and the acceptance time of different types of work orders provided in an embodiment of the present application;

[0058] Fig.13 This is an example diagram of the ratio relationship between the actual collection and the initial collection of different types of work orders provided in an embodiment of the present application;

[0059] Fig.14 This is an example diagram of the proportion of purchase applications of different types of work orders provided in the embodiment of the present application that are of the DM order type;

[0060] Fig.15 It is a kernel density diagram of spare parts requisition data and special type requisition ratio provided in the embodiment of the present application;

[0061] Fig.16 It is a flow chart of a method for determining requisition data based on a spare parts code set provided in an embodiment of the present application;

[0062] Fig.17 is a flow chart of a method for determining a spare parts code set provided in an embodiment of the present application;

[0063] Fig.18 is a schematic diagram of a data processing device provided in an embodiment of the present application;

[0064] Fig.19 It is a schematic diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0065] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0066] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0067] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0068] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0069] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0070] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0071] like Figure 1 As shown, the embodiment of the present application provides a system structure for implementing a data processing method, including a user terminal 100, a communication module 200 and a server 300. The user terminal 100 communicates with the server 300 through the communication module 200.

[0072] The user terminal 100 is deployed with an inventory strategy model tool 110 for spare parts. The tool 110 provides an operation interface 120 for determining inventory parameters based on spare parts requisition data. The spare parts may be nuclear power spare parts, etc. After the user inputs the target spare parts code through the operation interface 120, a data processing request carrying the target spare parts code is sent to the server 300. The target requisition data adapted to the target spare parts code is received from the server 300, and the inventory parameters adapted to the spare parts indicated by the target spare parts code are determined.

[0073] The server 300 is used to parse the data processing request sent by the user terminal, and obtain multiple pieces of initial requisition data matching the target spare part code within the specified statistical time period, as well as the material requisition method of each piece of initial requisition data, and perform data integration processing adapted to the material requisition method on the initial requisition data to obtain integrated requisition data; then, perform type identification adapted to the material requisition method on the integrated requisition data to obtain an identification result; if the identification result indicates that the integrated requisition data is interference requisition data, then remove the interference requisition data to obtain the target requisition data. Finally, the target requisition data is sent to the interface 120 of the user terminal 100 for display.

[0074] In the application, the user terminal 100 can be, but is not limited to, various electronic devices that can deploy inventory policy model tools, such as mobile phones, tablet computers, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptop computers, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA), handheld devices and other electronic devices. The embodiments of the present application do not impose any restrictions on the specific types of electronic devices.

[0075] In application, the communication module 200 can provide communication solutions including wireless local area networks (WLAN) (such as Wi-Fi networks), Bluetooth, Zigbee, mobile communication networks, global navigation satellite systems (GNSS), frequency modulation (FM), near field communication (NFC), infrared (IR) and the like applied to network devices. The communication module can be one or more devices integrating at least one communication processing module. The communication module 200 can include an antenna, which can have only one array element or an antenna array including multiple array elements. The communication module can receive electromagnetic waves through the antenna, frequency modulate and filter the electromagnetic wave signals, and send the processed signals to the processor. The communication module can also receive the signal to be sent from the processor, frequency modulate and amplify it, and convert it into electromagnetic waves for radiation through the antenna.

[0076] In the application, server 300 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), as well as big data and artificial intelligence platforms.

[0077] Before describing the data processing method of the embodiment of the present application, a process in which the server inputs the historical usage data of spare parts into the inventory strategy model to calculate the inventory parameters of the spare parts is described in detail.

[0078] like Figure 2 As shown, the inventory parameter configuration of nuclear power spare parts mainly includes three functions, namely, spare parts historical requisition data analysis and processing module (including requisition classification and duplicate code substitution integration, spare parts return data integration, special type requisition data identification), spare parts inventory parameter setting module (including spare parts basic data, spare parts classification, inventory strategy model parameters), spare parts inventory strategy model library module. Among them, special type requisition data is also called interference requisition data, which is invalid for inventory strategy model. The main function of spare parts historical data analysis and processing module is to process the historical requisition data of spare parts, including historical requisition data integration, identification and elimination of interference requisition data, etc.

[0079] In the application, the main function of the spare parts inventory parameter setting module is to classify spare parts based on the spare parts requisition data, and then determine the spare parts inventory model parameters. The main function of the spare parts inventory strategy model library module is to determine the corresponding inventory strategy model based on the model parameters of the spare parts inventory parameter setting module (usually, different types of inventory strategy models are used to calculate inventory parameters for different types of spare parts, such as the reorder point of spare parts with high requisition frequency is the "safety stock + requisition quantity within the procurement cycle" model, and the reorder point of spare parts with low requisition frequency is the "demand forecast based on probability distribution" model) and model setting parameters (such as the setting value of the availability rate), and calculate the spare parts library parameters in combination with the spare parts historical requisition data output by the spare parts historical data analysis and processing module.

[0080] In applications, the accuracy of historical spare parts requisition data directly affects the effectiveness of the inventory strategy model. As input information for the inventory strategy model, there are often abnormal situations in the requisition data, such as the return of spare parts. In order to meet the non-negative value requirements of the input information of the inventory strategy model, the data with negative requisition quantities during the statistical period are often directly corrected to 0. This method not only increases the overall requisition data volume of spare parts, but also distorts the probability distribution data of spare parts requisition based on the time dimension. In addition, since nuclear power plants will carry out maintenance activities such as transformation, replacement, and defect elimination, the probability of subsequent requisition of the same scale for spare parts requisitioned in such maintenance activities is extremely low. This type of requisition data for spare parts is a special type of requisition data that is invalid for the inventory strategy model, also known as interference requisition data.

[0081] The data processing method provided in the embodiment of the present application focuses on the specific implementation of the spare parts historical data analysis and processing function. The data processing process includes the data integration process of the returned material data and the identification process of the interfering collection data. By integrating the returned material data of the spare parts, the situation where the collection quantity is negative is avoided, and the spare parts collection data after data integration is consistent with the actual collection probability distribution. By identifying and marking the interfering collection data, the user can choose to automatically remove the interfering collection data. Since the processed collection data does not contain the interfering collection data, the inventory parameters calculated by the inventory policy model are more accurate, which can improve the accuracy of spare parts demand, optimize the inventory structure, control the inventory level, and also increase the proportion of users accepting the calculation results of the inventory policy model tool.

[0082] Next, the data processing method provided by the embodiment of the present application is described in detail. In one embodiment, Figure 3 As shown, a data processing method is provided, which is applied to a computer device (which may be Figure 1 The example of the server in FIG. 100 is used to illustrate the method, which includes the following steps S101 to S104:

[0083] Step S101, obtaining a plurality of pieces of initial requisition data matching the target spare part code within a specified statistical time period.

[0084] In the application, the target spare parts refer to the spare parts that need to be calculated for inventory parameters through the inventory policy model. The specified statistical time period refers to the selection of the statistical interval based on the granularity requirements of the requisition data required by the inventory model. Usually, statistics are performed on a yearly or monthly basis. Taking the data of the past five years as an example, there are 5 historical requisition data according to annual statistics and 60 historical requisition data according to monthly statistics. For any target spare part, the server obtains the initial requisition data within the specified time period according to its target spare part code, that is, the historical requisition data within the statistical period. The initial requisition data includes the return data for spare parts and the interference requisition data formed by the requisition of spare parts in historical maintenance activities. In order to eliminate the above-mentioned requisition data that is invalid for the inventory policy model, the server performs data integration processing on each initial requisition data to obtain the target requisition data that meets the input of the inventory policy model.

[0085] Step S102, obtaining the material collection method of the initial collection data, and performing data integration processing on the initial collection data adapted to the material collection method to obtain integrated collection data, wherein the collection quantity of the integrated collection data is a non-negative number.

[0086] In the application, spare parts are picked up by picking orders. The picking method in which a maintenance work order is attached to the picking order is called work order picking, and the picking method without a work order is called non-work order picking. Data integration processing refers to the data integration of the return data in the initial picking data, so that the picking quantity of the obtained integrated picking data is a non-negative number, so as to meet the requirement that the picking quantity is a non-negative number when the integrated picking data is used as the input of the subsequent inventory strategy model. For the initial picking data generated by different picking methods, the server adopts data integration processing adapted to the picking method, pre-processes the initial picking data, and obtains the integrated picking data.

[0087] In the application, the server identifies the integrated collection data to determine whether the integrated collection data is interference collection data.

[0088] Step S103, identifying the type of the integrated requisition data that is compatible with the material requisition method to obtain an identification result.

[0089] In the application, the collection quantity of the integrated collection data is a non-negative number. Although it meets the requirement that the collection quantity is a non-negative number when it is input to the inventory policy model, the integrated collection data may be interference collection data. Therefore, the server identifies the integrated collection data through an identification method adapted to the collection method, and obtains an identification result that characterizes whether the integrated collection data is interference collection data. If the identification result characterizes that the integrated collection data is not interference collection data, it can be used as the input of the inventory policy model. If the identification result characterizes that the integrated collection data is interference collection data, it cannot be used as the input of the inventory policy model.

[0090] In the application, for the integrated requisition data of non-work order requisition, since it is difficult to verify the installation and use of spare parts after requisition, all non-work order requisitions are usually marked as interference requisition data or non-interference requisition data based on the management regulations of the power plant. For example, for the inventory strategy setting of important spare parts, since the shortage of important spare parts has a great impact on the safe production of nuclear power plants, the main purpose is to ensure on-site supply, and non-work order requisitions are set as non-interference requisition data, that is, the requisition data of non-work order requisition is also used to calculate inventory parameters. For the inventory strategy setting of non-important spare parts (which can be combined with the purchase unit price of spare parts), since the shortage of non-important spare parts has a small impact on the safe production of nuclear power plants, the main purpose is to control the inventory of power plants, and non-work order requisitions are set as interference requisition data, that is, non-work order requisition data is not used to calculate inventory parameters.

[0091] In the application, for the integrated requisition data of the material requisition method of work order requisition, the server comprehensively judges whether the requisition data interferes with the requisition data based on the keyword information of the attached maintenance work order, the work order type, the relationship between the spare parts requisition time and the acceptance time, the proportion of spare parts single requisition in the specified time and total historical requisition, the proportion of DM orders in the purchase application, the proportion of spare parts single requisition in the specified time and total historical requisition, the number of spare parts requisitions within the specified time, etc. DM orders refer to manually submitted inventory material purchase applications, which correspond to inventory material purchase applications (DR orders) triggered by the material requirement planning (Master Production Schedule, MRP) module.

[0092] Step S104: if the recognition result indicates that the integrated collection data is interference collection data, the interference collection data is removed to obtain target collection data.

[0093] In the application, the server removes the interfering collection data from the integrated collection data to obtain the target collection data that can be input into the inventory strategy model. If the target collection data is linked to the maintenance order number, then one maintenance order number corresponds to one target collection data.

[0094] In applications, such as Figure 1The operation interface 120 provided by the client deployment inventory strategy model tool connected to the server is shown in the figure. In this interface, the target spare parts code is filled in to obtain the target requisition data from the server and classify the target requisition data. For example, the requisition data of spare parts in the past five years is automatically obtained based on the statistical dimension of the year, and is divided into annual sudden requisition data, annual first planned requisition data (the spare parts demand time and spare parts demand quantity of this type are determined, which can also be called annual planned A-type requisition data), and annual second planned requisition data (the spare parts demand time and spare parts demand quantity of this type are determined, which can also be called annual planned B-type requisition data).

[0095] In the application, when calculating the historical annual use data of spare parts, the detailed work order use data of spare parts can be sorted out so that users can identify and mark special types of work order use data. The inventory strategy model system tool includes all work order use data in the calculation scope of the inventory strategy model by default. If some work order use data belongs to interference use data, users can uncheck such work order use data, and such work order use data will not participate in the subsequent inventory strategy model calculation.

[0096] In the application, a function of removing special types of requisition data with one click is added to the system tool. After the user selects it, the system tool can automatically remove the interfering requisition data.

[0097] Based on the content of this application, the spare parts requisition data can be automatically classified, the duplicate code and replacement data can be integrated, and the return data can be integrated. The historical requisition classification data of the spare parts and the requisition data of special types of work orders can be obtained. The system tool can automatically calculate the recommended inventory parameters based on this type of requisition data, such as Figure 1 In the interface 120 shown in the figure, the data processing method of the present application can be operated through the inventory strategy model system tool to automatically calculate the historical use data of spare parts, including historical use data integration, identification and elimination of abnormal use data, and construction of an annual use data matrix. On this basis, users can adjust the spare parts use data included in the calculation to assist in setting the nuclear power spare parts inventory strategy and improve the efficiency and quality of inventory parameter setting.

[0098] In the application, the server receives a data processing request sent by the inventory strategy model system tool, parses the target spare parts code in the data processing request, obtains the historical requisition data of the target spare parts code within the specified statistical time period, and uses a data integration method that is adapted to the requisition method for the historical requisition data to eliminate interfering requisition data and obtain target requisition data, which is used to assist in setting the nuclear power spare parts inventory strategy and improve the efficiency and quality of inventory parameter setting.

[0099] In this embodiment, by focusing on the collection data that matches the specific target spare parts code within the specified statistical time period, the pertinence and accuracy of the analysis data are ensured, and the interference of irrelevant data is avoided. The data is adaptively integrated and processed according to different collection methods, so that the data format is unified, which is convenient for subsequent analysis and processing, and improves the efficiency and accuracy of data processing. Through type recognition technology, it is possible to automatically identify which collection data may be interference data, such as erroneous entry, duplicate records or abnormal collection. This step is crucial for data cleaning and can significantly improve data quality. After identifying and removing interfering collection data, the target collection data set obtained is purer, which can effectively improve the efficiency and accuracy of spare parts collection data analysis. It can then effectively assist in setting the nuclear power spare parts inventory strategy, improve the efficiency and quality of inventory parameter settings, and increase the proportion of users accepting the calculation results of the inventory strategy model tool.

[0100] In one embodiment, Figure 4 As shown, when the material picking method is work order picking, the implementation method of step S102 includes the following steps S201 to S203:

[0101] Step S201, obtaining at least one maintenance work order number attached to the initial collection data.

[0102] In the application, each piece of requisition data corresponds to a material requisition form. The material requisition form records the material requisition form number and the number of requisitions (or returns) for spare parts. In the case where the material requisition method is work order requisition, each piece of requisition data is also associated with a maintenance work order number. The server sorts out the corresponding maintenance work order number based on the material requisition form number. Multiple pieces of initial requisition data may be associated with one maintenance work order number or multiple different maintenance work order numbers.

[0103] Step S202: determining a first collection data group corresponding to each maintenance work order number from a plurality of initial collection data.

[0104] In the application, the server uses the maintenance work order number as an index to group multiple initial requisition data to obtain the first requisition data group corresponding to each maintenance work order number. The first requisition data group includes all requisition data corresponding to the corresponding maintenance work order number. The attribute information of the requisition data in the first requisition data group includes the requisition order number, maintenance work order number, requisition (returned) quantity, maintenance work order start date, etc.

[0105] Step S203, perform data integration on each of the first collection data groups to obtain integrated collection data corresponding to each of the maintenance work order numbers; wherein the integrated collection data includes a target collection quantity and a target collection time for spare parts, the target collection quantity being the difference between the total collection quantity for spare parts and the total return quantity for spare parts in the first collection data group, and the collection time being the first collection time for spare parts in the maintenance work order.

[0106] In the application, the server integrates the collection data in each first collection data group, and can obtain an integrated collection data corresponding to each maintenance work order number. In the application, since there may be multiple collection and return records for spare parts in the same maintenance work order, integration is required, and each maintenance work order corresponds to only one collection data. For example, if the first collection data group corresponding to a maintenance work order includes: 3 collections in February 2020, 1 collection in April 2020, and 2 returns in May 2020, then after the collection data of the maintenance work order is integrated, the collection of 2 in February 2020 will be retained (the collection time in the integrated collection data is the first collection time of the maintenance work order, and the collection quantity is all the collection quantities for spare parts in the work order minus all the return quantities, that is, 3+1-2). The integrated collection data includes the maintenance work order number, the total collection quantity, and the start date of the maintenance work order. Normally, the total quantity of a maintenance work order should be greater than or equal to 0. If the total quantity is less than 0 (for example, the user returns spare parts that were collected for other work orders or non-work orders under this maintenance work order, which is inconsistent with current management regulations and is an abnormal situation), the quantity will be directly assigned a value of 0.

[0107] In one embodiment, Figure 5 As shown, when the material picking method includes non-work order material picking, the implementation method of step S102 includes the following steps S301 to S303:

[0108] Step S301, obtaining at least one batch number associated with the initial requisition data, wherein the batch number is set when the spare parts are inspected and accepted.

[0109] In the application, the batch number will be set when the spare parts are accepted. Usually, the batch numbers of the spare parts in the same batch of the same purchase order will be the same after acceptance. When the material collection method is non-work order collection, the server will filter out the initial collection data that is not linked to the maintenance work order number, and count the one or more batch numbers included in the filtered initial collection data.

[0110] Step S302, determining a second collection data group corresponding to each batch number from the plurality of initial collection data.

[0111] In the application, the server uses the batch number as an index to search for the second collection data group corresponding to each batch number from the filtered initial collection data. The attribute information of the collection data in the second collection data group includes the collection order number, batch number, collection (return) quantity, and collection (return) time.

[0112] Step S303, for each second requisition data group, sort them in reverse order according to the material return time, starting from the first requisition data after sorting, and fill the returned materials for the subsequent requisition data in turn until the requisition quantity for the spare parts is a non-negative number, to obtain an integrated second requisition data group, and determine each of the integrated second requisition data groups as the integrated requisition data.

[0113] In the application, since there are multiple records of requisition and return of spare parts in the same batch number, the server integrates the data of each second requisition data group to obtain an integrated second requisition data group. The specific process of data integration for the second requisition data group is: starting from the latest time of return discovery, the return material is filled to the previous requisition data until the requisition quantity is non-negative.

[0114] For example, it is assumed that there are non-work order pickings for equipment parts in two batch numbers, and the picking data is shown in the following Table 1:

[0115] Table 1

[0116]

[0117] For the data of batch 1, the latest material return time is January 2, 2022, and 2 materials are returned. Therefore, the returned materials are filled forward, and the initial collection quantity on July 3, 2021 is 6. After the returned materials are filled, the collection quantity is corrected to 4; 3 materials are returned on May 2, 2020, and the returned materials are filled forward. On February 15, 2020, 2 materials are collected. After the correction, there is still 1 returned material. The returned materials are filled forward, and the collection on February 3, 2020 is corrected to 3. Therefore, the effective collection of batch 1 is: 3 on February 3, 2020, 4 on July 3, 2021, and 3 on May 1, 2023. Similarly, the effective collection of batch 2 can be calculated as: 3 on September 2, 2021, 2 on November 1, 2022, and 1 on May 9, 2023. On August 15, 2021, two material returns occurred, and the returns were carried forward. On July 13, 2021, one was issued, and the issued quantity was still -1. The reason is that the effective issuance time is earlier than the statistical interval (for example, the statistical interval is 2020-2024, and one spare part of the returned material was issued before 2020). At this time, the issued quantity on July 13, 2021 is set to 0.

[0118] In this embodiment, by tracing the batch number of the spare parts when they enter the warehouse, the initial collection data can be accurately grouped to ensure that each group of data corresponds to a specific batch of spare parts. Then, the return time is sorted in reverse order, and the most recent return record is processed first. This "first in, last out" strategy helps to update the inventory status in real time and ensure the accuracy of the inventory data. In the process of filling the collection data with returned materials, starting from the latest collection record, it is gradually adjusted forward until the requirement of non-negative spare parts collection quantity is met. This not only solves the problem of abnormal inventory quantity caused by returned materials, but also improves inventory turnover efficiency, reduces unnecessary inventory backlogs, and ultimately achieves refined and efficient inventory management.

[0119] In one embodiment, Figure 6 As shown, the material picking method is work order picking, the integrated picking data is linked to the maintenance work order number, and the implementation method of step S103 includes the following steps S401 to S404:

[0120] Step S401, obtaining a first keyword and a second keyword for identifying interference data, wherein the first keyword has a higher accuracy rate in identifying interference data than the second keyword.

[0121] In the application, the maintenance work orders of the integrated collection data of the work order collection method have description information, which is used to introduce the main content of the maintenance work order. The average length of the work order description is about 24 characters. By extracting the keyword information in the work order description, it can be used to assist in identifying the interference collection data.

[0122] Based on the identified multiple interference claim data, the corresponding work order description information is sorted out, and the keywords with the highest tf-idf value ranking are calculated according to the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm. The calculation method of the tf-idf value is shown in formula (1):

[0123] tf-idf=tf×idf (1)

[0124] Among them, tf is the frequency of occurrence of a keyword, that is, word frequency, and idf is the inverse document probability of a keyword.

[0125] Treat all the work order descriptions of the maintenance work order as one document and automatically split the keywords of all the documents. Calculate the word frequency value of a keyword according to formula (2):

[0126]

[0127] Among them, n 1 is the number of times the keyword appears, n 2is the total number of words in the document.

[0128] For each keyword, the inverse document probability is calculated according to formula (3):

[0129]

[0130] Among them, n 3 is the number of document collections (open source document collection data can be used), n 4 It is the number of texts in the document collection that contain the keyword, where 1 is added to avoid the denominator being zero.

[0131] Taking the keyword information of a nuclear power plant's work order description as an example, the top 100 keywords with the tf-idf value are calculated. The accuracy, recall and F value are as follows: Figure 7 shown. Figure 7 In the figure, the horizontal axis is the accuracy, the vertical axis is the recall, and the diameter of the bubble is the F of each keyword. β The value is a comprehensive evaluation value of accuracy and recall. Among them, accuracy refers to the accuracy of using the keyword to judge whether the work order collection belongs to the interference collection data. The calculation formula of accuracy is as follows:

[0132]

[0133] Among them, n 5 To use this keyword for screening, the number of interference work order items containing this keyword in the sample; n 6 To use this keyword for filtering, the number of work order items that contain this keyword is sampled.

[0134] In applications, the recall rate of keywords is also a key factor for evaluating the effect of keyword recognition. The recall rate represents the proportion of the keyword in the total sample. The calculation formula (5) of the recall rate is as follows:

[0135]

[0136] Among them, n 5 To use this keyword for screening, the number of items in the sample containing this keyword that interfere with the work order; n 7 is the number of all interference work order items in the sample. In the application, the comprehensive evaluation value F of accuracy and recall is also used. β Comprehensively evaluate the judgment effect of keywords, F β The value is calculated as follows:

[0137]

[0138] Wherein, β is a weighted value. The larger the β value is, the greater the weight of the recall rate is, and the smaller the β value is, the greater the weight of the accuracy rate is. In this embodiment, the accuracy rate is more important than the recall rate, and the β value is set to 0.15 in the application.

[0139] In the application, the interference collection data is identified based on keywords, and the keywords are sorted from high to low according to the accuracy corresponding to the keywords. Keywords can be divided into three categories based on the aforementioned key accuracy. Keywords with an accuracy greater than 90% (a set value, which can be adjusted according to management needs) are the first keywords of the first type, keywords with an accuracy between 40% and 90% (a set value, which can be adjusted according to management needs) are the second keywords of the second type, and keywords with an accuracy lower than 40% (a set value, which can be adjusted according to management needs) are the third keywords of the third type. For example, Figure 8 As shown in the figure, taking the description key of the work order of a nuclear power plant as an example, the first keywords belonging to the first type generally include: transformation, aging, replacement, expansion inspection, experience feedback, governance, elimination of defects, special projects, and increase. The second keywords belonging to the second type generally include: improvement, problem feedback, redundancy, cable laying, reinforcement, addition, non-conformity, optimization, replacement, total ticket, tail item, concentration, etc. The third keywords belonging to the third type generally include: change, defect, investigation, crack, new type, stirring, etc. In the application, due to the low accuracy of the third keyword of the third type, the first keyword of the first type and the second keyword of the second type are usually selected to identify the interference data to improve the recognition accuracy.

[0140] Step S402, obtaining description keywords of the maintenance work order attached to the integrated requisition data.

[0141] In the application, for each integrated collection data, the server reads the description keywords of the maintenance work order to which it is attached. Based on the keyword recognition method, the integrated collection data with the collection method of the work order is identified. Keyword recognition refers to identifying the interference collection data through the first keyword and the second keyword. Among them, the accuracy of the first keyword in identifying the interference collection data is higher than that of the second keyword.

[0142] Step S403: determining the association relationship between the description keyword, the first keyword and the second keyword.

[0143] In the application, the association relationship between the description keyword and the first keyword and the second keyword includes: the description keyword includes the first keyword, the description keyword does not include the first keyword but includes the second keyword, and the description keyword does not include the first keyword and does not include the second keyword.

[0144] Step S404: Based on the association relationship, the integrated collection data is type identified to obtain an identification result.

[0145] In the application, if the association relationship is that the description keyword includes the first keyword, then all the collection data of this work order are interference collection data; if the association relationship is that the description keyword does not include the first keyword but includes the second keyword, it is necessary to use the support vector machine (SVM) model to calculate and determine whether the work order collection is interference collection data; if the association relationship is that the description keyword does not include the first keyword and does not include the second keyword, then all the collection data of this work order are non-interference collection data.

[0146] In this embodiment, by using the first keyword and the second keyword with different recognition accuracy rates, the system can evaluate the reliability of the data in a hierarchical manner. First, the first keyword with a high accuracy rate is used to quickly screen out obviously worthless requisition records, which greatly improves processing efficiency. Subsequently, the nature of the remaining data can be judged more carefully by combining the association between the description keyword of the maintenance work order and the two keywords. This method not only reduces the negative impact of invalid data on the analysis results, but also ensures the retention of important information. The final recognition results are more accurate and reliable, providing a solid foundation for subsequent data analysis, inventory management decisions and resource optimization.

[0147] In one embodiment, Fig. 9 As shown, the implementation of step S404 includes the following steps S501 to S503:

[0148] Step S501: If the association relationship indicates that the description keyword includes the first keyword, a recognition result is obtained that the integrated collection data is interference collection data.

[0149] Step S502: If the association relationship represents that the description keyword does not include the first keyword but includes the second keyword, the integrated requisition data is type identified through a type identification model to obtain an identification result, and the identification result indicates whether the integrated requisition data is interference requisition data.

[0150] Step S503: If the association relationship indicates that the description keyword does not include the first keyword and the second keyword, a recognition result is obtained that the integrated collection data is non-interference collection data.

[0151] In the application, if the keywords describing the relationship representation include the first keyword, all the collection data of the work order are interference collection data; if the keywords describing the relationship representation do not include the first keyword but include the second keyword, it is necessary to calculate through the type recognition model to determine whether the collection of the work order is interference collection data; if the keywords describing the relationship representation do not include the first keyword and do not include the second keyword, all the collection data of the work order are non-interference collection data.

[0152] In this embodiment, an efficient and accurate mechanism is provided to identify and filter interference data in the collection data. Through the dual judgment of the pre-set keyword rules and the type recognition model, the system can automatically distinguish the validity of the data. First, if the collection data contains known interference keywords (first keywords), it is directly judged as interference data; secondly, for data containing another type of keywords (second keywords), it is further verified by a more sophisticated type recognition model to ensure that non-interference data will not be misjudged; finally, when the data contains neither interference nor keywords that require additional judgment, the system defaults to valid data. Such a design not only reduces the burden of manual review and improves the speed of data processing, but also enhances data quality control through multi-level recognition strategies, which has a significant positive impact on optimizing data analysis, decision making and subsequent process automation.

[0153] In one embodiment, the type recognition model is a binary classification model, which is trained based on the collection data samples. The labels of the collection data samples represent whether they are interference collection data, and the sample features of the collection data samples are obtained by normalizing the collection data.

[0154] In the application, for the collection data of the material collection method of work order, there are only two identification results for whether the collection data is interference collection data, which is interference collection data and not interference collection data. Therefore, the problem can also be converted into a two-category classification problem. K-Nearest Neighbors (KNN), support vector machine, random forest, neural network and other algorithms can be used to handle classification problems.

[0155] Based on this, the server can use an artificial intelligence-based type recognition model to perform type recognition on the integrated historical collection data to obtain a recognition result, which is used to characterize whether the integrated collection data is interference collection data. The type recognition model is essentially a binary classification model, so the type recognition model can be implemented through the SVM model.

[0156] In application, SVM is a classifier based on data learning in the field of machine learning. It can find a classification hyperplane from limited training samples to maximize the interval between classification surfaces, so that the classification prediction of unknown samples can be most accurate. Therefore, the support vector machine algorithm can be used to classify the impact of sub-component failure. The algorithm model based on support vector machine is a black box model. The nonlinear mapping relationship between the input and output of the model is realized by the support vector machine. Therefore, the model can be built and applied as follows: First, extract the spare parts sample data set, and randomly sample 2 / 3 of the special type of spare parts requisition data identified manually as training samples, and the remaining 1 / 3 as test samples. Secondly, preprocess the spare parts sample data set and normalize the data sets of discrete attributes and numerical attributes. Finally, build and train the support vector machine model.

[0157] The specific process of extracting the collection data sample is explained. The collection data sample is obtained based on the collection data. For each collection data, the keyword information of the work order, the work order type, the relationship between the spare parts collection time and the acceptance time, the ratio of actual collection to initial collection, the proportion of DM orders in the purchase application, the proportion of single collection of spare parts in the total collection of history within a specified time, the number of collections of spare parts within a specified time and other specified attribute data are selected as the sample features of the collection data sample, and whether the collection data is interference collection data is used as the sample label of the collection data sample.

[0158] In the application, in order to improve the accuracy of inventory parameters, the server preprocesses the specified attribute data in the requisition data (such as normalization, etc.). For discrete attribute data, the discrete type data is normalized to 0-1 by establishing rules; for numerical attribute data, the linear transformation method is used to normalize the data to 0-1. Since some data in the database are incorrect or the values ​​are too large, affecting the normalization effect, the maximum and minimum values ​​during data normalization can be optimized.

[0159] The construction method of the type recognition model is explained. In one embodiment, an objective function corresponding to the type recognition model is constructed, a hyperplane is used to divide the spare parts requisition data into special types, the intervals between different fault categories are calculated, the conditions of the hyperplane when the interval is maximum are analyzed, and the optimal hyperplane is calculated.

[0160] Assume there are n samples of data {(x 1 ,y 1 ),(x 2 ,y 2 ),...,(x n ,y n )}, where x i is the normalized work order type I 1 1. The relationship between spare parts receiving time and acceptance time2 , the ratio of actual use to initial use I 3 , DM order ratio in purchase application 4 , the coefficient of total number of times spare parts are used in a single use within a specified time I 5 , the value of the single-time use ratio coefficient of spare parts within a specified time I 6 ,y i is the label of the sample of the collected data, which is used to indicate whether the collected data belongs to the interference collected data. By constructing a hyperplane classifier, as shown in formula (7),

[0161] f(x)=sign(w T x+b) (7)

[0162] Among them, sign(·) is the sign function, w is the normal vector of the hyperplane, representing the direction of the hyperplane, and w T is the transposed vector of the normal vector, and b is the intercept of the hyperplane, which represents the distance between the hyperplane and the origin.

[0163] In the application, the hyperplane H 0 :y=w T +b=0 and two parallel and equidistant hyperplanes H 1 :y=w T +b=1 and H 2 :y=w T +b=-1,hyperplane H 1 and the hyperplane H 2 Therefore, the problem of finding the optimal hyperplane is a problem of finding the optimal value with inequality constraints, which can be expressed as sty i (w t x+b)≥1, where y i (w t x+b)≥1 is an inequality constraint. Since some samples cannot be separated by a linear hyperplane classifier, a slack variable ξ can be introduced i When dealing with inseparable samples, the problem of finding the optimal hyperplane can be expressed as formula (8):

[0164]

[0165] Among them, C is a regularization parameter, which is used to control the penalty for misclassified samples. A larger C value means a smaller tolerance for misclassification and a tendency to choose a larger interval, which may lead to overfitting; a smaller C value is the opposite. i (w t x+b)≥1-ξ i ,ξ i ≥0 is an inequality constraint.

[0166] In the application, the Lagrange multiplier method is used to solve formula (8), where the Lagrange multiplier calculation formula is shown in formula (9):

[0167]

[0168] Among them, α i is the Lagrange multiplier, used to ensure the equivalence of the original problem. i is the Lagrange multiplier used for the transformation of constraints.

[0169] In the application, by changing w, b, ξ in formula (9) i ,α,β, and then we can get the expression shown in formula (10):

[0170]

[0171] In applications, nonlinear problems can be transformed into linear problems in a high-dimensional space through nonlinear transformation, and an appropriate inner product function k(x 1 ,x 2 ) replaces Φ(x i )Φ(x j ), that is, the optimization function above can be transformed into formula (11),

[0172]

[0173] The classification function that can be obtained is The inner product function k(x 1 ,x 2 ) is also called the kernel function. In this application, the Gaussian kernel function is used to train the model, that is,

[0174] In the above SVM model, the output of the model can be optimized by adjusting the C parameter and the σ parameter. Let the accuracy of the training sample output by the SVM model be p 1 , the accuracy of the test sample output by SVM is p 2 , based on the accuracy of the training samples and the test samples, set the comprehensive accuracy, as shown in Formula 12,

[0175]

[0176] In the application, when the training accuracy and test accuracy are both high and close to each other, the comprehensive accuracy is high, because the comprehensive accuracy p value output by SVM is maximized by continuously adjusting the C parameter and σ parameter. Since the C parameter and σ parameter must be positive numbers, the parameter setting problem can be transformed into a constrained multivariable optimization problem, and the optimal C parameter and σ parameter can be calculated by using particle swarm optimization (PSO), genetic algorithm (GA), etc.

[0177] In the application, taking the genetic algorithm as an example, the specific process of setting the C parameter and the σ parameter is further explained. The genetic algorithm is a global search heuristic algorithm that simulates the genetic mode of selection, crossover, and mutation in nature, and adopts the evolutionary idea of ​​survival of the fittest to solve the optimization problem. The genetic algorithm maps the solution of the problem to be solved to the chromosome in the coding space through the chromosome encoding method. The chromosome represents the individual in the group. The collection of chromosomes is called the group. The genetic algorithm starts from a certain initial group and uses genetic operators to act on the group to generate the next generation of groups. According to the fitness function, relatively excellent individuals are selected from these chromosomes, and iterative genetic operations are performed on these individuals to generate a new generation of populations. This process is repeated until the set convergence conditions are met.

[0178] The related genetic algorithm can quickly explore the global search space, but the mining effect of the search space is not good. It is usually manifested as the genetic algorithm approaches a certain optimal value relatively quickly, and then encounters a slow sorting problem, approaching the optimal value very slowly. In addition, in the process of selecting the next generation of the related genetic algorithm, the selection probability of individuals is highly correlated with the fitness value, which is prone to premature phenomenon. The genes of all individuals in the population converge to the local optimal value, and it is difficult to jump out of the local optimal solution. When encountering complex multi-peak optimization problems, the optimization accuracy is not high.

[0179] In one embodiment, Fig.10 As shown, in order to solve the problem of slow convergence of genetic algorithms and easy to fall into local optimal values, the genetic algorithm provided in the embodiment of the present application can expand the search space of individuals by adding random factors when individuals are in a slow convergence state or have fallen into a local optimal value, and reset the positions of some individuals based on the current optimal position of the population. The reset probability changes dynamically. In the early stage of the algorithm iteration, the reset probability is small, so that individuals will not search around the current optimal value too early in the early stage, causing local convergence; in the middle and late stages of the algorithm, the reset probability is large, which can make individuals search more near the current optimal value and speed up the convergence. The corresponding pseudo code snippet is as follows: {

[0180] S1: Initialization. Set the number of iterations = 1, for individuals in the population:

[0181] S1.1: Randomly initialize individual decimal positions Where i is the individual sample and j is the dimension;

[0182] S1.2: Convert individual decimal places Convert to binary position

[0183] S2: Repeat the iterative process until the number of iterations reaches the set m times, or the deviation between the optimal fitness value and the target value D is less than e;

[0184] S2.1: Calculate the fitness value of the individual based on the individual decimal position:

[0185] S2.2: Select the individual with the best fitness value among all individuals and record its decimal position as S best , whose fitness value is recorded as g best ;

[0186] S2.3: Based on the fitness value of the individual, use the roulette algorithm to select the individual to be inherited to the next generation and establish the parent generation library;

[0187] S2.4: Based on the parent pool, select the parents in turn, randomly select the current individual as the mother, and use the single-point crossover method to produce offspring.

[0188] Step 2.5: For the individuals of the offspring, based on the set mutation probability, use the reversal mutation method to update the offspring

[0189] S2.5: Individual binary positions Convert to decimal position

[0190] S2.6: Randomly select one offspring individual, let Ensure that the optimal fitness value of the offspring of the population is not inferior to that of the parent generation;

[0191] S2.7: If the conditions for the new random factor are met, execute S2.7.1 to S2.7.4, otherwise go directly to S2.8;

[0192] S2.7.1: In In , 4 individuals are randomly selected and their positions are reassigned;

[0193] S2.7.2: Construct a chaotic array sequence r, where r(n+1)=4r(n)[1-r(n)], r(1)∈(0,1) and

[0194] S2.7.3: For individual 1, assign its position to S best+ k(r-0.5), where k = 2(mt) / m;

[0195] S2.7.4: For individual 2, assign its position to S best (kr-0.5k+1), where k=2(mt) / m;

[0196] S2.7.5: For individual 3, assign its position to [S best +k(r-0.5)+s best (kr-0.5k+1)] / 2, where k=2(m-) / m:

[0197] S2.7.6: For individual 4, assign its position to S best , to prevent the historical best position from being covered in the process of adding new factors;

[0198] S2.8: Convert individual decimal positions Convert to binary position

[0199] S2.9: Update the number of iterations, set t = t + 1;

[0200] If t≥m, or |g best -d|≤e, go to step 3.1, otherwise return to step 2.1;

[0201] S3: Output calculation results:

[0202] S3.1: Output the optimal position S best , and the optimal fitness value g best .

[0203] }

[0204] Among them, in S2.7, by calculating the fitness values ​​of all individuals in the population, when When the condition is met, the step of adjusting the positions of some individuals is performed, where f2 is the median fitness value of all individuals in the population, f1 is the optimal fitness value of all individuals in the population, and p is a parameter that controls the probability of triggering the new random factor, which can be taken as 0.2. In the early stage of the algorithm iteration, since the median fitness value of all individuals in the population is significantly different from the optimal fitness value, the probability of triggering the new random factor is low; in the middle and late stages of the algorithm iteration, the median fitness value of all individuals in the population is slightly different from the optimal fitness value, and the probability of triggering the new random factor is high. In S2.7.2, is a chaotic array sequence, where r(n) is the value of the chaotic sequence obtained by the nth iteration, which is distributed in the interval (0,1). Based on the chaotic array and the optimal position of the current population, the positions of some individuals are reassigned, as shown in S2.7.3, S2.7.4, and S2.7.5.

[0205] In the application, by constructing a fitness value evaluation model, the genetic algorithm is used to calculate the optimal C parameter and σ parameter. The fitness value calculation formula is shown in formula 13:

[0206]

[0207] Where p is the comprehensive accuracy value output by the SVM model in formula (12). By performing a reciprocal operation on it, the target maximization can be converted into the target minimization. In order to avoid the error of dividing by 0 due to the denominator being 0, p+0.1 is used instead of p in the denominator of formula (13) to ensure that the denominator will not be 0.

[0208] For the SVM model, when a certain C parameter and σ parameter are assigned to the model, the accuracy of the training sample calculated based on the SVM model is p 1 , the accuracy of the test sample is p 2 , the comprehensive accuracy p value is calculated based on formula (12), and the fitness value of the genetic algorithm is calculated based on formula (13) to find the optimal C parameter and σ parameter of the SVM model.

[0209] In this embodiment, the type recognition model is trained by a genetic algorithm, which can improve the training accuracy of the type recognition model and improve the accuracy of interference data identification using the category recognition model.

[0210] In one embodiment, Fig.11 As shown, the implementation method of determining the target attribute value includes the following steps S601 to S603:

[0211] Step S601, obtaining the attribute type of each attribute of the requisition data.

[0212] In the application, the server normalizes the specified attributes in the collection data used as the characteristics of the collection sample, and the obtained normalized attribute data is used as the characteristics of the collection sample. Among them, different normalization methods are used to implement the normalization operation for the specified attributes of different attribute types. The server determines the attribute type of each specified attribute. The attribute type may include at least discrete attributes (such as attributes whose attribute values ​​are enumerated attributes), numerical attributes, etc.

[0213] Step S602: If the attribute type is a discrete attribute, the discrete attribute value is normalized based on a preset normalization condition to obtain a target attribute value.

[0214] In the application, for data whose attribute type is a discrete attribute, the discrete type data is normalized to a target attribute value of 0 to 1 based on a preset normalization condition.

[0215] In the application, the specified attributes read from the requisition data include the work order type, the relationship between the spare parts requisition time and the acceptance time, the ratio of actual requisition to initial requisition, the proportion of DM orders in the purchase application, the proportion of single requisition of spare parts in the total historical requisition within the specified time, the number of requisitions of spare parts within the specified time, etc. Among them, the data with discrete attributes of the work order type is represented by an enumeration type, and is usually divided into planned A-type requisition, planned B-type requisition, and sudden requisition by the work order requisition. That is, the work order requisition type is the data with discrete attributes. The requisition of maintenance spare parts for NMO type is planned requisition, and the requisition of maintenance spare parts for non-NMO type is sudden requisition; for planned requisition, the MRP / PR field in the maintenance standard package corresponding to the maintenance work order is empty (that is, when it is automatically attached to the maintenance work order, the reservation type is immediate), the corresponding maintenance project spare parts are Class A spare parts (planned Class A requisition), and the spare parts of other preventive maintenance projects are Class B spare parts (planned Class B spare parts). By analyzing the interference collection data (or special collection data) in the sample database, it is found that the proportion of special collection data in planned Class A collection is the lowest, and the proportion of special collection data in sudden collection is the highest. It is necessary to normalize the data of planned Class A collection, planned Class B collection, and sudden collection to 0-1, so the planned Class A collection can be set to 0, the planned Class B collection can be set to 0.5, and the sudden collection can be set to 1. The normalization formula is shown in the following formula (14):

[0216]

[0217] Step S603: If the attribute type is a numerical attribute, a linear transformation is used to normalize the attribute value to obtain a target attribute value.

[0218] In the application, for data whose attribute type is a numerical attribute, a linear transformation method is used to normalize the data to a target attribute value of 0 to 1.

[0219] The normalization operation of the ratio of spare parts collection time to acceptance time is explained. In the application, for special work order collection data, procurement is usually triggered in advance before the collection demand occurs, and the spare parts are collected and installed on site after arrival. Therefore, the value of the collection time minus the acceptance time is small, that is, the spare parts are stored in the warehouse for a short time. By analyzing the relationship between the collection time and acceptance time of different types of work order collection data, the spare parts collected for special types of work orders are stored in the warehouse for about 75% of the storage time of regular spare parts. The distribution relationship is as follows: Fig.12 Therefore, the difference between the spare parts collection time and the acceptance time can be used as the input parameter I of the SVM model. 2 .

[0220] Count the mean of the spare parts collection time and acceptance time of all the same types of spare parts under the work order, and normalize it with the median value of the work order of this type to obtain the normalized value of the spare parts collection time and acceptance time. The calculation formula is as follows:

[0221]

[0222] Among them, t 2 is the median value of this type of work order, t 1 is the average of the corresponding type of use under the work order. When t 1 Value greater than 2 times t 2 When t 2 For example, under a work order, 5 spare parts were collected, of which 2 were planned Class A collections (spare parts collection time-acceptance time were 3 months and 5 months respectively), and 3 were planned Class B collections (spare parts collection time-acceptance time were 11 months, 5 months, and 13 months respectively). Under this work order, the mean time of planned Class A collections was 4 months, and the mean time of planned Class B collections was 14.5 months. Assuming that the median value of planned Class A collections is 14.7, and the median value of planned Class B collections is 12.8, the I of the 2 planned Class A collections under this work order is 2 The value is 0.86 ((2*14.7-4) / (2*14.7)). The I of the three planned B-type spare parts under this work order is 2 The value is 0.43((2*12.8-14.5) / (2*12.8)).

[0223] The normalization operation of the ratio of actual collection to initial collection is explained. In the application, for the collection data of special work orders, the proportion of returned materials is usually low, that is, the ratio of actual collection to initial collection is high. Fig.12 As shown in the figure, the ratio of actual collection to initial collection under different types of work orders is quite different. For planned A-type work order demand, the ratio of actual collection to initial collection is higher, that is, the proportion of returned materials is lower; for sudden work order demand, the ratio of actual collection to initial collection is lower, that is, the proportion of returned materials is higher. Therefore, the ratio of actual collection to initial collection can be used as the input parameter I of the SVM model. 3 .

[0224] In the application, the average value of the ratio of actual use to initial use of all spare parts of the same type under the work order is calculated, and it is normalized with the median value of the use of this type of work order to obtain the normalized value of the ratio of actual use to initial use. The calculation formula is as follows:

[0225]

[0226] Among them, p 2 is the median value of this type of work order, p 1 is the mean of the corresponding type of requisition under the work order. For example, there are 5 spare parts requisitions under a work order, of which 2 are planned type A requisitions (the actual requisitions are 5 and 3 respectively, and the initial requisitions are 5 and 9), and 3 are planned type B requisitions (the actual requisitions are 3, 2, and 8 respectively, and the initial requisitions are 5, 2, and 9). Under this work order, the mean ratio of the planned type A requisition to the actual requisition is 0.667 (0.5*(1 / 1+3 / 9)), and the mean ratio of the planned type B requisition to the initial requisition is 0.829 (0.333*(3 / 5+2 / 2+8 / 9)). Based on Fig.13 It can be obtained that the median of the ratio of the actual use of planned A parts to the initial use is 95.9%, and the median of the ratio of the actual use of planned B parts to the initial use is 91.8%. Therefore, the I 3 The value is 0.35 (0.667 / 0.959 / 2). The I 3 The value is 0.45 (0.829 / 0.918 / 2).

[0227] The normalization operation of the DM order ratio in the purchase application is explained. For the special type of work order requisition data, the probability of directly requisitioning inventory materials is usually low, that is, based on the demand for such special type of spare parts, it is highly likely that future demand will be manually formulated, and the purchase application (DM order) will be manually triggered to purchase spare parts, and then the spare parts will be installed on site after they arrive. The proportion of purchase applications for different types of work orders that are DM orders is as follows: Fig.14 As shown in Figure 1, the DM proportion of special type work order collection data is significantly higher than the DM proportion of all work order collection data. Therefore, the DM proportion of work order collection data can be used as the input parameter I of the SVM model. 4 .

[0228] Count the average of the proportion of all spare parts DM orders of the same type under the work order, and normalize it with the average value of the work order of this type to obtain the normalized value of the ratio of actual use to initial use. The calculation formula is as follows:

[0229]

[0230] Among them, p 4 is the median value of the proportion of DM orders of this type of work order, p 3 is the average proportion of DM orders of the corresponding type under this work order. 3 Value greater than 2 times p 4 When it is corrected to 2 times p 4. Based on the work order, the corresponding material requisition is sorted out. Based on the batch number of the material requisition, the corresponding purchase order PO is sorted out. Based on the purchase order PO, the corresponding purchase requisition PR is sorted out, and then the purchase requisition is obtained as a DM order or a DR order. For example, under a work order, 5 spare parts are collected, 2 of which are planned Class A collections (1 spare part is a DM order and 1 is a DR order), and 3 are planned Class B collections (2 of which are DM orders and 1 is a DR order). Under this work order, the proportion of DM orders for planned Class A collections is 0.5 (1 / 2), and the proportion of DM orders for planned Class B collections is 0.333 (1 / 3). Based on Fig.14 It can be obtained that the median of the ratio of the actual use of planned A parts to the initial use is 38.8%, and the median of the ratio of the actual use of planned B parts to the initial use is 52.2%. Therefore, the I 4 The value is 0.59 ((1+0.5-2*0.388) / (2-2*0.388)). The I 4 The value is 0.32 (0.333 / 0.522 / 2).

[0231] The preprocessing method for the total number of times of use coefficient and the single number of uses ratio coefficient is explained. In the application, by analyzing the total number of times spare parts are used in the work order and the ratio of this use in the total use, a kernel density diagram of spare parts use data and special type use ratio is constructed, such as Fig.15 As shown in the figure, the darker the color, the higher the probability that it belongs to a special type of collection data. Because in the sample data, different combinations of total collection times and single collection quantity ratios have different sample sizes, and the sample size of some combinations is lower than the threshold number (for example, 5 samples), the data of this type of combination is not included. Fig.15 There are some blank spaces in the upper right corner (there is no sample data, so the probability of it belonging to the special type of data is assigned to 0). Fig.15From the probability distribution, we can see that when the total number of withdrawals is small and the proportion of single withdrawals is high, the probability of it belonging to the special type of withdrawal is high; as the total number of withdrawals increases, when the single withdrawal data is between 20% and 60%, it also has a high probability of belonging to the special type of withdrawal data. In short, the relationship between the total number of withdrawals coefficient, the single withdrawal proportion coefficient and the special type of withdrawal proportion is relatively complex and cannot be calculated by a simple linear fitting method. The relationship between the three can only be described by probability distribution. For example, when the total number of receipts in the past five years is 5 times, the proportion of the number of receipts this time is 60% of the total receipts, and the probability that it is a special type of work order receipt is 13.3%; when the total number of receipts in the past five years is 7 times, the proportion of the number of receipts this time is 40% of the total receipts, and the probability that it is a special type of work order receipt is 16.2%; when the total number of receipts in the past five years is 15 times, the proportion of the number of receipts this time is 20% of the total receipts, and the probability that it is a special type of work order receipt is 7.1%. It is difficult to directly use a linear function to describe the relationship between these three.

[0232] Based on the above information, we can get the number of spare parts picked within a specified time, the proportion of a single spare part pick in the total historical pick within a specified time, and whether the work order pick belongs to a special type of pick quantity. There is a certain correlation between them. Therefore, the total pick frequency coefficient and the single pick quantity ratio coefficient can be used as the input parameters of the SVM model. 5 with I 6 .

[0233] The calculation method of the total number of times of use coefficient is shown in formula 18.

[0234]

[0235] Among them, k 1 k is the number of times the spare part is used within a specified time range (such as the data for the past five years). 2 is the set threshold (such as 20 times) so that I 5 Normalized to 0~1.

[0236] The calculation method of the coefficient value of the single-time quantity ratio is shown in Formula 19.

[0237]

[0238] Among them, n 8 is the number of spare parts used this time, n 9 It is the total number of spare parts used in a specified time range (such as statistics for the past five years).

[0239] In this embodiment, by obtaining the attribute type of each attribute of the data, and performing corresponding normalization processing according to the attribute type, different types of attribute values ​​can be effectively converted into a unified standard, thereby improving the efficiency and accuracy of data analysis and processing. For discrete attributes, normalization is performed using preset normalization conditions to ensure that the weights of each attribute value in the calculation are relatively fair, avoiding excessive impact of certain large or small values ​​on the results. For numerical attributes, linear transformation is used for normalization so that all attribute values ​​fall within the same range, which is convenient for subsequent data analysis and comparison.

[0240] In one embodiment, Fig.16 As shown, the implementation of step S101 includes the following steps S701 to S702:

[0241] Step S701, based on the target spare part code, in the order of material collection time, the merged first code and the second code before replacement are traced back step by step to obtain a spare part code set, wherein the spare part code set includes the target spare part code, the first code and the second code.

[0242] In the application, when the server obtains the spare parts requisition data, in addition to the spare parts code of the target spare parts, it can also use the spare parts code set associated with the target spare parts. The spare parts code set includes other codes associated with operations such as duplicate code integration and code replacement in the maintenance activities linked to the target spare parts.

[0243] In the application, nuclear power maintenance activities can be divided into two categories: overhaul and daily maintenance. Both overhaul and daily maintenance include planned maintenance and emergency maintenance. According to the above maintenance activity classification, the spare parts historical collection data can be divided into: overhaul planned A-type collection data, overhaul planned B-type collection data, overhaul emergency collection data, overhaul non-work order collection data, daily planned A-type collection data, daily planned B-type collection data, daily emergency collection data, and daily non-work order collection data. Spare parts collected for overhaul planned A-type requirements are spare parts that must be replaced during overhaul according to the plan; spare parts collected for overhaul planned B-type requirements are spare parts that are selectively replaced according to the status of on-site equipment during overhaul according to the plan; spare parts collected for overhaul emergency requirements are spare parts that are not included in the plan but need to be temporarily replaced due to emergencies during overhaul; spare parts collected for overhaul non-work order refer to spare parts that are directly collected without issuing a work order. If the above four types of spare parts collection data are changed from overhaul to daily maintenance, the spare parts collection data will correspond to spare parts collected for daily planned Class A demand, spare parts collected for daily planned Class B demand, spare parts collected for daily emergency demand, and spare parts collected for daily non-work order demand. Since nuclear power plant spare parts have duplicate code mergers and item substitutions, it is necessary to integrate the spare parts collection data involving substitutions, that is, to integrate the original coded collection data into the new coded collection data, so the new coded collection data includes the new coded and original coded collection data. In addition to replacing spare parts due to failure of on-site equipment or preventing failure of on-site equipment, there are also some special situations that may cause spare parts replacement, such as aging replacement, modification replacement, replacement, defect elimination replacement, etc. If the spare parts inventory parameters guarantee this type of spare parts demand, there is a high possibility of inventory backlog. For example, the average annual usage of a certain spare part in a power plant is 1, and the procurement cycle is 1 year. According to the usage calculation in the procurement cycle, 1 minimum inventory is reserved. In a certain year, 21 spare parts are required for aging treatment. At this time, according to the usage calculation in the procurement cycle, 5 minimum inventories are reserved. The average annual usage of this spare part is 1, so 4 more stocks are reserved every year, resulting in inventory backlog. Therefore, when calculating the inventory parameters of spare parts, the usage data of spare parts for aging treatment should be excluded.

[0244] like Fig.17 As shown, where x 1 For a given spare part code, trace back the merged duplicate code and the code before replacement based on the code. For example, x 2 For x 1 The encoding before substitution; x 3 For x 2 The repeated code, and x 3 There is no other duplicate code, and no code before replacement; x 4 For x 2 The encoding before substitution; x 5 、x6 、x 7 For x 4 Replace the previous encoding, and x 5 、x 6 、x 7 There is no other duplicate code, and no code before replacement; x 8 For x 2 The repeated code, and x 8 There is no other duplicate code, and no code before replacement; x 9 For x 2 The encoding before substitution; x 10 For x 9 Replace the previous encoding, and x 10 There is no other duplicate code, nor the code before replacement. For spare part x 1 After integrating duplicate codes and replacement data, the spare parts code set {x 1 、x 2 、x 3 、x 4 、x 5 、x 6 、x 7 、x 8 、x 9 、x 10}.

[0245] Step S702: Based on the spare part code set, a plurality of initial requisition data matching the target spare part code within a specified statistical time period are obtained.

[0246] In the application, the server sorts out the historical work orders and material requisition data corresponding to all spare parts in the code set. If the user chooses not to integrate duplicate codes and replacement data, only code x 1 The corresponding historical work order is the material picking order data.

[0247] In this embodiment, by tracing back the changes in spare parts codes, including the target spare parts code, the first code (i.e., the code before being merged) and the second code (i.e., the code before being replaced), a complete spare parts code set can be constructed. This helps to fully understand the evolution of spare parts and ensure the continuity and integrity of the data, especially when the spare parts codes are changed or integrated. Based on the constructed spare parts code set, the collection data within the specified statistical time period can be obtained more accurately. This means that even if the spare parts code has changed, it can ensure that all relevant collection records are included in the analysis, avoiding data omissions or distortions caused by code changes.

[0248] The present application also provides a data processing device for executing the steps in the above data processing method embodiment. The data processing device may be a computer device (such as Figure 1A virtual appliance in a server in a computer system is run by a processor of a computer device, or it may be the computer device itself.

[0249] like Fig.18 As shown, the data processing device 400 provided in the embodiment of the present application includes:

[0250] The acquisition module 410 is used to acquire a plurality of pieces of initial requisition data matching the target spare part code within a specified statistical time period.

[0251] The data integration module 420 is used to obtain the material collection method of the initial collection data, and perform data integration processing on the initial collection data that is adapted to the material collection method to obtain integrated collection data, wherein the collection quantity of the integrated collection data is a non-negative number.

[0252] The identification module 430 is used to identify the type of the integrated requisition data that is compatible with the material requisition method to obtain an identification result.

[0253] The removal module 440 is configured to remove the interfering collection data to obtain the target collection data if the recognition result indicates that the integrated collection data is interfering collection data.

[0254] In one embodiment, the material collection method is work order material collection, and the data integration module is also used to obtain at least one maintenance work order number attached to the initial collection data; determine the first collection data group corresponding to each maintenance work order number from multiple initial collection data; perform data integration on each of the first collection data groups to obtain integrated collection data corresponding to each maintenance work order number; wherein the integrated collection data includes a target collection quantity and a target collection time for spare parts, the target collection quantity is the difference between the total collection quantity for spare parts in the first collection data group and the total material return quantity for spare parts, and the collection time is the first collection time for spare parts in the maintenance work order.

[0255] In one embodiment, the material collection method includes non-work order material collection, and the data integration module is also used to obtain at least one batch number associated with the initial collection data, and the batch number is set when the spare parts are accepted for storage; from multiple initial collection data, determine the second collection data group corresponding to each batch number; for each second collection data group, sort them in reverse order according to the return time, starting from the first collection data after sorting, and fill the return materials for the subsequent collection data in turn until the collection quantity for the spare parts is non-negative, so as to obtain an integrated second collection data group, and determine each integrated second collection data group as the integrated collection data.

[0256] In one embodiment, the material collection method is work order collection, and the integrated collection data is attached with a maintenance work order number. The identification module is also used to obtain a first keyword and a second keyword for identifying interference collection data, and the accuracy of the first keyword in identifying interference collection data is higher than that of the second keyword; obtain the description keyword of the maintenance work order attached with the integrated collection data; determine the association relationship between the description keyword, the first keyword and the second keyword; based on the association relationship, perform type identification on the integrated collection data to obtain an identification result.

[0257] In one embodiment, the identification module is also used to obtain an identification result that the integrated collection data is interfering collection data if the association relationship represents that the description keywords include the first keyword; if the association relationship represents that the description keywords do not include the first keyword but include the second keyword, then through a type identification model, type identification is performed on the integrated collection data to obtain an identification result, and the identification result indicates whether the integrated collection data is interfering collection data; if the association relationship represents that the description keywords do not include the first keyword and do not include the second keyword, then a identification result that the integrated collection data is non-interfering collection data is obtained.

[0258] In one embodiment, the type recognition model used in the recognition module is a binary classification model, which is trained based on the collection data sample. The label of the collection data sample represents whether it is interference collection data, and the sample characteristics of the collection data sample are obtained by normalizing the collection data.

[0259] In one embodiment, the identification module is also used to obtain the attribute type of each attribute of the usage data; if the attribute type is a discrete attribute, the discrete attribute value is normalized based on a preset normalization condition to obtain a target attribute value; if the attribute type is a numerical attribute, a linear transformation method is used to normalize the attribute value to obtain a target attribute value.

[0260] In one embodiment, the acquisition module is also used to trace back the merged first code and the second code before replacement step by step according to the order of material collection time based on the target spare part code to obtain a spare part code set, wherein the spare part code set includes the target spare part code, the first code and the second code; based on the spare part code set, obtain multiple initial collection data matching the target spare part code within a specified statistical time period.

[0261] In applications, each module in the data processing device may be a software program module, or may be implemented by different logic circuits integrated in a processor, or may be implemented by multiple distributed processors.

[0262] Fig.19 The following is a schematic diagram of the structure of a computer device for implementing a data processing method provided in an embodiment of the present application. Fig.19 As shown, the computer device 1000 of this embodiment includes: at least one processor 1010 ( Fig.19 Only one is shown in the figure) a processor, a memory 1020, and a computer program 1030 stored in the memory 1020 and executable on the at least one processor 1010, and when the processor 1010 executes the computer program 1020, the steps in any of the above-mentioned data processing method embodiments are implemented.

[0263] The computer device 1000 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The data processing device / terminal device may include, but is not limited to, a processor 1010 and a memory 1020. Those skilled in the art will appreciate that Fig.17 The computer device 1000 is merely an example and does not constitute a limitation on the computer device 1000 . The computer device 1000 may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the computer device 1000 may also include input and output devices, network access devices, etc.

[0264] The processor 1010 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0265] In some embodiments, the memory 1020 may be an internal storage unit of the computer device 1000, such as a hard disk or memory of the computer device 1000. In other embodiments, the memory 1020 may also be an external storage device of the computer device 1000, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 1000. Further, the memory 1020 may also include both an internal storage unit of the computer device 1000 and an external storage device. The memory 1020 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory 1020 may also be used to temporarily store data that has been output or is to be output.

[0266] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0267] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0268] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0269] An embodiment of the present application provides a computer program product. When the computer program product runs on a mobile terminal, the mobile terminal can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0270] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electric carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0271] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0272] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example 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 performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0273] In the embodiments provided in the present application, it should be understood that the disclosed devices / network equipment and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are merely schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0274] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0275] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A data processing method, characterized in that: include: Obtain multiple pieces of initial requisition data matching the target spare part code within the specified statistical time period; Acquire the material collection method of the initial collection data, and perform data integration processing on the initial collection data adapted to the material collection method to obtain integrated collection data, wherein the collection quantity of the integrated collection data is a non-negative number; Performing type identification on the integrated requisition data that is compatible with the requisition method to obtain an identification result; If the identification result indicates that the integrated collection data is interference collection data, the interference collection data is removed to obtain target collection data.

2. The data processing method according to claim 1, characterized in that: The material collection method is work order material collection, and the initial collection data is subjected to data integration processing adapted to the material collection method to obtain integrated collection data, including: Obtain at least one maintenance work order number attached to the initial collection data; Determine a first collection data group corresponding to each maintenance work order number from the plurality of initial collection data; Data integration is performed on each of the first collection data groups to obtain integrated collection data corresponding to each of the maintenance work order numbers; wherein the integrated collection data includes a target collection quantity and a target collection time for the spare parts, the target collection quantity is the difference between the total collection quantity for the spare parts in the first collection data group and the total return quantity for the spare parts, and the collection time is the first collection time for the spare parts in the maintenance work order.

3. The data processing method according to claim 1, characterized in that: The material picking method includes non-work order material picking, and the initial picking data is subjected to data integration processing adapted to the material picking method to obtain integrated picking data, including: Acquire at least one batch number associated with the initial requisition data, where the batch number is set when the spare parts are accepted for storage; Determine a second collection data group corresponding to each batch number from the plurality of initial collection data; For each second requisition data group, it is sorted in reverse order according to the material return time. Starting from the first requisition data after sorting, the subsequent requisition data are filled with returned materials in turn until the requisition quantity for spare parts is a non-negative number, and an integrated second requisition data group is obtained, and each integrated second requisition data group is determined as the integrated requisition data.

4. The data processing method according to claim 1, characterized in that: The material collection method is work order collection, and the integrated collection data is linked to the maintenance work order number. The step of performing type identification on the integrated requisition data that is compatible with the requisition method to obtain an identification result includes: Acquire a first keyword and a second keyword for identifying the interference data, wherein the first keyword has a higher accuracy rate in identifying the interference data than the second keyword; Obtaining description keywords of the maintenance work order to which the integrated requisition data is attached; Determine an association relationship between the description keyword, the first keyword, and the second keyword; Based on the association relationship, the integrated collection data is identified by type to obtain an identification result.

5. The data processing method according to claim 4, characterized in that: Based on the association relationship, the integrated use data is type identified to obtain an identification result, including: If the association relationship indicates that the description keyword includes the first keyword, then an identification result is obtained that the integrated collection data is interference collection data; If the association relationship represents that the description keyword does not include the first keyword but includes the second keyword, then type identification is performed on the integrated requisition data through a type identification model to obtain an identification result, and the identification result indicates whether the integrated requisition data is interference requisition data; If the association relationship represents that the description keyword does not include the first keyword and does not include the second keyword, then an identification result is obtained that the integrated collection data is non-interference collection data.

6. The data processing method according to claim 5, characterized in that: The type recognition model is a binary classification model, which is obtained by training based on the collection data samples. The labels of the collection data samples represent whether they are interference collection data, and the sample features of the collection data samples are obtained by normalizing the collection data.

7. The data processing method according to claim 6, characterized in that: The method further comprises: Get the attribute type of each attribute of the requisition data; If the attribute type is a discrete attribute, the discrete attribute value is normalized based on a preset normalization condition to obtain a target attribute value; If the attribute type is a numerical attribute, a linear transformation method is used to normalize the attribute value to obtain a target attribute value.

8. The data processing method according to any one of claims 1 to 7, characterized in that: The obtaining of multiple pieces of initial requisition data matching the target spare part code within the specified statistical time period includes: Based on the target spare part code, the merged first code and the second code before replacement are traced back step by step in the order of material collection time to obtain a spare part code set, wherein the spare part code set includes the target spare part code, the first code and the second code; Based on the spare part code set, a plurality of initial requisition data matching the target spare part code within a specified statistical time period are obtained.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the data processing method according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the data processing method according to any one of claims 1 to 8 is implemented.

Images