A Big Data-Based Intelligent Management Method and System for the Industrial Chain

Abstract

This invention discloses a big data-based intelligent management method and system for the industrial supply chain, specifically relating to the field of industrial supply chain supervision and management. It addresses the challenges of identifying implicit outsourcing processes, locating shared outsourcing nodes, and quantifying outsourcing scale in existing industrial chain management processes. By acquiring the bill of materials and waste recycling records corresponding to the order delivery cycle, a process-waste output mapping table is established, corresponding to process nodes and theoretically produced waste categories. This identifies waste missing categories and, combined with the cross-process waste collaborative fluctuation relationship between adjacent process nodes, determines the suspected outsourcing process location. Furthermore, based on the transport vehicle driving records, non-registered stopping points and load change sequences are extracted to identify shared outsourcing nodes. The process-waste output mapping table is then used to calculate the outsourcing volume of processes, achieving spatial traceability and scale estimation of implicit outsourcing processes, thereby improving the precision of industrial chain collaborative supervision.

Description

Technical Field

[0001] This invention relates to the field of industrial supply chain supervision and management technology, and more specifically, to a big data-based intelligent management method and system for the industrial chain. Background Technology

[0002] As the division of labor in the industrial chain becomes increasingly refined, manufacturing companies often outsource some processes to different suppliers to improve production efficiency and reduce manufacturing costs during actual order fulfillment. In some manufacturing scenarios, to cope with insufficient equipment capacity, tight order cycles, or lack of specific processing capabilities, suppliers may transfer some processes to unregistered processing points.

[0003] Because such outsourcing activities typically occur outside the registered production address of manufacturing enterprises, and the relevant processing nodes are not included in the formal supply chain management system, problems such as concealed processing, illegal subcontracting, and gray capacity sharing are easily formed. Existing supply chain management methods mainly rely on order flow records, manual inspections, or information self-reported by enterprises for supervision, making it difficult to effectively verify actual processing activities. Especially in the waste recycling stage, there are usually relatively stable process correlations between the types of waste, waste output rates, and waste transfer behaviors corresponding to different processes. However, existing technologies lack the comprehensive analytical capabilities to analyze the collaborative relationships between waste processes, waste absence behaviors, and the correlation characteristics between non-registered and unregistered nodes, making it impossible to identify actual hidden outsourcing processes from waste recycling activities.

[0004] Furthermore, existing solutions lack effective means to identify the behavior of multiple suppliers sharing the same unregistered processing node, making it difficult to spatially locate and estimate the scale of shared outsourced capacity hidden behind waste transfer behavior. This leads to problems in supply chain management, such as difficulty in identifying outsourced processes, difficulty in quantifying the scale of outsourced processes, and difficulty in tracing hidden processing nodes. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a big data-based intelligent management method and system for the industrial chain to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A big data-based intelligent management method for the industrial chain includes the following steps: S1. Obtain the bill of materials between the manufacturer and the target supplier, as well as the weighing ledger in the waste recycling records, within the order delivery cycle; S2. Using the bill of materials as the decomposition object, break down the material input of each process node along the product manufacturing process. Based on historical production records, establish a mapping relationship between the process node and the output rate corresponding to the theoretical output waste category, and form a process-waste output mapping table. S3. Aggregate the weighing ledger by waste category into the waste composition distribution of the order delivery cycle, and align it with the process-waste output mapping table to extract the missing waste categories. S4. Based on historical production records, calculate the collaborative fluctuation range corresponding to the waste output rate of each adjacent process node, and obtain the process pointers of the missing categories of waste output rate that exceed the cross-process collaborative fluctuation range, as suspected outsourcing process positions. S5. Using the suspected outsourced work station as an index, extract the coordinates of non-registered dwelling points and the corresponding load change sequence from the transportation vehicle driving records of waste recycling records within the current order delivery cycle; S6. Retrieve the waste recycling records of similar suppliers of the target supplier during their order delivery cycle, count the proportion of valid outsourcing events corresponding to non-registered outsourcing point coordinates, and identify shared outsourcing nodes. S7. Based on the net increase in load of transfer vehicles in the shared outsourcing nodes, calculate the outsourcing quantity of the process in combination with the process-waste output mapping table, and output the coordinates of the shared outsourcing nodes, the outsourcing quantity of the process, and the set of related similar suppliers.

[0008] As a further aspect of the present invention, in step S1, obtaining the bill of materials between the manufacturing enterprise and the target supplier, as well as the weighing ledger in the waste recycling record, specifically includes: The bill of materials is generated within the delivery cycle after the manufacturing enterprise places an order with the target supplier. The list entries record the material codes, order identifiers and corresponding material quantities that constitute the product production. Retrieve the weighing ledger entries within the order delivery cycle corresponding to the current order. The weighing ledger entries record the waste material name, order identifier, and corresponding transfer vehicle identifier. The driving records of the corresponding transfer vehicle during the order delivery period are retrieved based on the vehicle identification. The driving records include the vehicle location trajectory, timestamp, and vehicle load record.

[0009] As a further aspect of the present invention, in step S2, forming a process-waste output mapping table specifically includes: Based on the material consumption relationships marked at each process node in the product manufacturing process of the manufacturing enterprise, the materials in the bill of materials are allocated to the corresponding process nodes according to the material code and the material quota input quantity; Based on the processing type of each process node, the names of the waste materials generated by the corresponding process node are extracted from the product's historical production records. After adding a unique process code prefix to the waste material names, they are used as the theoretical output waste material categories for the corresponding process node. In the product's historical production records, the ratio sequence of waste weight to material quota input for each waste category in each production batch is statistically analyzed. The mean of the ratio sequence is extracted as the output rate of the corresponding waste category, and a process-waste output mapping table is established.

[0010] As a further aspect of the present invention, in step S3, the missing categories of extracted waste specifically include: Read the waste categories recorded in each entry of the weighing ledger corresponding to the current order identifier, and classify and collect all weighing ledger entries within the order delivery cycle based on the waste categories to establish the waste composition distribution. Iterate through the total actual weight of each waste category in the waste composition distribution, and combine it with the output rate of the corresponding waste category in the process-waste output mapping table to calculate the theoretical waste output weight. Mark the waste categories whose total actual weight is lower than the lower bound of the theoretical waste output weight as missing categories.

[0011] As a further aspect of the present invention, in step S4, obtaining the process direction for missing product categories that exceed the cross-process waste collaborative fluctuation range specifically includes: From the historical order delivery cycles set before the current order delivery cycle, retrieve the total actual weight of each type of waste produced by adjacent process nodes in the process-waste output mapping table within each historical delivery cycle. Calculate the ratio sequence of the total actual weight of the two types of waste for each pair of adjacent process nodes. The minimum and maximum values ​​of the ratio sequence constitute the cross-process waste collaborative fluctuation range of the adjacent process node pair. Based on the process-waste output mapping table, read the first process node corresponding to the missing product category and the second process node that has a process adjacency relationship with the first process node, extract the actual total weight of waste products of the first process node and the second process node in the current order delivery cycle, and calculate the adjacency ratio. If the adjacent ratios do not fall within the cross-process waste collaborative fluctuation range corresponding to the first process node and the second process node, then the first process node corresponding to the missing product category will be marked as a suspected outsourcing process position.

[0012] As a further aspect of the present invention, step S5, extracting the coordinates of non-registered dwelling points whose dwell time exceeds the preset dwelling determination and the corresponding load change sequence, specifically includes: Read the theoretical waste category name corresponding to the suspected outsourced process position in the process-waste output mapping table, retrieve the ledger entry from the weighing ledger by category name and order identifier, and obtain the driving record of the transfer vehicle; Extract the trajectory segment from the start of the current order delivery cycle to the last weighing timestamp in the driving record of the transfer vehicle, extract the stopping positions where the dwell time exceeds the preset dwell judgment threshold from the trajectory segment, and use the location coordinates of the stopping positions as the dwell point coordinates; The coordinates of the stopping points that are consistent with the production address registered by the manufacturer are removed, the coordinates of the non-registered stopping points are retained, and the on-board weighing readings that are consistent with the time period of the non-registered stopping points are extracted from the driving records to form a load change sequence.

[0013] As a further aspect of the present invention, in step S6, identifying the shared external node specifically includes: Screen similar suppliers at the same material supply level as the target supplier, retrieve the waste recycling records corresponding to the orders placed by the production enterprise to these similar suppliers within a set historical period, and extract the stationing events and load change sequences of the transfer vehicles at the coordinates of non-registered stationing points. Filter out the dwell events in the load change sequence where the direction of change is consistent and the waste category name is consistent with the theoretical output waste category name of the suspected outsourced process station, and record such dwell events as valid outsourced dwell events. The proportion of valid outsourcing events among the total events corresponding to non-registered outsourcing point coordinates is counted. If the proportion exceeds the preset sharing judgment ratio, the corresponding non-registered outsourcing point coordinates are marked as shared outsourcing nodes.

[0014] As a further aspect of the present invention, in step S7, calculating the outsourcing quantity of the process specifically includes: Extract valid outsourcing stay events from shared outsourcing nodes, and sum the net increase in the load capacity of the transfer vehicle corresponding to each valid outsourcing stay event according to the waste type to obtain the total weight of outsourced waste. Read the output rate of the suspected outsourced process position from the process-waste output mapping table, divide the total weight of outsourced waste by the output rate to obtain the outsourced quantity of the process, and output the location coordinates of the shared outsourced node, the outsourced quantity of the process, and the identifier of the associated similar supplier corresponding to the shared outsourced node.

[0015] On the other hand, the present invention provides a big data-based intelligent management system for the industrial chain, comprising: The data acquisition module is used to acquire the bill of materials and the weighing ledger in the waste recycling records corresponding to the order delivery cycle between the production enterprise and the target supplier, and associate the order identifier, material code, transfer vehicle identifier and transfer vehicle driving record to establish the correspondence between orders, waste and transfer vehicles. The process waste analysis module is used to establish a mapping relationship between the output rate of each process node and the theoretical output waste category based on the material consumption relationship and historical production records of each process node in the product manufacturing process, form a process-waste output mapping table, and identify missing categories that are below the lower limit of the theoretical waste output weight. The outsourcing identification module is used to calculate the cross-process waste coordination fluctuation range between adjacent process nodes based on historical production records, and to perform coordination fluctuation analysis on the total actual weight of waste received between the process node corresponding to the missing product category and adjacent process nodes to identify suspected outsourcing process positions. The shared outsourcing traceability module is used to extract the load change sequence corresponding to non-registered depots, count the proportion of effective outsourcing depot events of transfer vehicles from different suppliers of the same type at non-registered depots, identify shared outsourcing nodes, and calculate the outsourcing quantity of processes in combination with the process-waste output mapping table.

[0016] The technical effects and advantages of the intelligent supply chain management method and system based on big data of this invention are as follows: This invention establishes a production rate mapping relationship between process nodes and theoretically produced waste categories, and performs process correlation analysis on the waste composition distribution within the order delivery cycle. This enables automatic identification of abnormal waste absence behavior during production. Furthermore, by utilizing the cross-process waste coordination and fluctuation relationship between adjacent process nodes, it accurately locates suspected outsourced process positions, improving the reliability of implicit outsourced process identification. After identifying suspected outsourced process positions, this invention combines the driving records of transport vehicles, the coordinates of unregistered stopping points, and the load change sequence to perform correlation analysis on abnormal waste transport behavior, achieving the identification and spatial tracing of shared outsourced nodes. This effectively discovers the hidden processing behavior of multiple similar suppliers sharing the same unregistered processing node. Simultaneously, by combining the process-waste output mapping table to back-calculate the outsourced quantity of the process, it can quantify the scale of outsourced processing, improve the ability to identify, trace, and assess the scale of implicit outsourced behavior in the supply chain management process, thereby enhancing the precision of supply chain collaborative supervision. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a big data-based intelligent management method for the industrial chain according to the present invention; Figure 2 This is a schematic diagram of the structure of a big data-based intelligent management system for the industrial chain according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1 Figure 1 This invention presents a big data-based intelligent supply chain management method, which includes the following steps: S1. Obtain the bill of materials between the manufacturer and the target supplier, as well as the weighing ledger in the waste recycling records, within the order delivery cycle; S2. Using the bill of materials as the decomposition object, break down the material input of each process node along the product manufacturing process. Based on historical production records, establish a mapping relationship between the process node and the output rate corresponding to the theoretical output waste category, and form a process-waste output mapping table. S3. Aggregate the weighing ledger by waste category into the waste composition distribution of the order delivery cycle, and align it with the process-waste output mapping table to extract the missing waste categories. S4. Based on historical production records, calculate the collaborative fluctuation range corresponding to the waste output rate of each adjacent process node, and obtain the process pointers of the missing categories of waste output rate that exceed the cross-process collaborative fluctuation range, as suspected outsourcing process positions. S5. Using the suspected outsourced work station as an index, extract the coordinates of non-registered dwelling points and the corresponding load change sequence from the transportation vehicle driving records of waste recycling records within the current order delivery cycle; S6. Retrieve the waste recycling records of similar suppliers of the target supplier during their order delivery cycle, count the proportion of valid outsourcing events corresponding to non-registered outsourcing point coordinates, and identify shared outsourcing nodes. S7. Based on the net increase in load of transfer vehicles in the shared outsourcing nodes, calculate the outsourcing quantity of the process in combination with the process-waste output mapping table, and output the coordinates of the shared outsourcing nodes, the outsourcing quantity of the process, and the set of related similar suppliers.

[0020] In S1, the bill of materials between the manufacturer and the target supplier, as well as the weighing ledger in the waste recycling record, are obtained during the order delivery cycle.

[0021] The bill of materials (BOM) is generated simultaneously by the manufacturing company when it formally places a production order with the target supplier. Based on the product model, batch number, and production quantity corresponding to the order, the manufacturing company retrieves the standard material composition information of the corresponding product from its internal product process database and generates corresponding material entries according to the categories of raw materials actually processed during the product manufacturing process. Each material entry records a unique material code, order identifier, and corresponding material quantity. The material code is generated using the unified coding rules in the company's existing material management system; different metal sheets, plastic granules, electronic components, and auxiliary materials each have independent material codes. The order identifier uses the order serial number automatically assigned when the order is generated; all material entries within the same order delivery cycle are bound to the same order identifier. The corresponding material quantity is calculated based on the planned production quantity of the order and the product's standard material consumption quota. For example, if a single product requires 2.5 kg of aluminum sheet material and the order plans to produce 2000 units, the corresponding aluminum sheet material entry will be recorded as 5000 kg. After the order is formally placed, the manufacturing company stores the BOM in the order management record and maintains consistency between the order identifier and subsequent waste recycling records. To avoid confusion regarding material attribution between different orders, when an order delivery cycle changes, the valid time range under the corresponding order identifier is updated synchronously. Only waste recycling records generated within the corresponding order delivery cycle are allowed to participate in subsequent correlation analysis. For multiple orders executed concurrently by the same production enterprise within the same time period, a dual association method using order identifiers and material codes is used for differentiation, ensuring that waste from different orders does not overlap in subsequent waste composition distribution statistics. For orders with split production, the split sub-orders continue to inherit the original order identifier, and a split batch number is appended to the material entry.

[0022] When retrieving weighbridge entries for the current order's delivery cycle, the start and end times of the delivery cycle are determined based on the order identifier. Then, weighbridge entries within this time range are selected from the waste recycling records maintained by the waste recycler. After each waste transport, the waste recycler weighs the transfer vehicle on a fixed weighbridge and generates a corresponding weighbridge entry. Each weighbridge entry records the waste name, order identifier, and corresponding transfer vehicle identifier. The waste name is filled in according to the waste classification rules followed by the waste recycler during the receiving process; for example, aluminum scraps, copper powder scraps, and stainless steel stamping scrap each have independent waste names. The order identifier is simultaneously entered into the waste transfer manifest by the manufacturing company when registering the waste leaving the factory and is entered into the corresponding weighbridge entry after the waste recycler receives the waste, thus establishing a link between waste records and orders. The transfer vehicle identifier uses the vehicle registration number or license plate number as unique identification information, maintaining a unique correspondence for the same vehicle throughout the entire order delivery cycle. After completing the screening of the weighbridge records, the driving records of the corresponding transfer vehicles within the order delivery cycle are retrieved based on the vehicle identification. During transportation, the transfer vehicles continuously upload vehicle location trajectories, timestamps, and vehicle load records. The vehicle location trajectories are periodically collected by a satellite positioning terminal installed on the vehicle, with a fixed collection interval of 30 seconds. The vehicle uploads its current location coordinates every 30 seconds during operation. The timestamps use a standardized time format to record the corresponding trajectory collection time. The vehicle load records are collected in real-time by an onboard weighing device installed at the vehicle's axle. The onboard weighing device records the total vehicle load value at fixed time intervals, for example, every 60 seconds, and simultaneously writes the corresponding timestamp. To avoid load data fluctuations caused by vehicle vibration when the vehicle is unloaded, when the vehicle load is below 500 kg, the corresponding load record is uniformly corrected to the unloaded state value.

[0023] In step S2, a process-waste output mapping table is formed.

[0024] The complete manufacturing process records for the corresponding product are retrieved from the manufacturing enterprise. These records sequentially record the name of each process node, the processing content of each process, and the types of materials involved in the processing, according to the actual processing order of the product. The material consumption relationships are calibrated using the long-term, stable process formula records implemented by the manufacturing enterprise, with different material input rules corresponding to different process nodes. For example, in the production scenario of metal structural parts, the stamping process node corresponds to the consumption relationship of aluminum plate raw material cutting, the welding process node corresponds to the consumption relationship of welding wire and shielding gas, and the surface treatment process node corresponds to the consumption relationship of cleaning fluid and spraying materials. During the process finalization stage, the manufacturing enterprise generates standard material input records for each process node through statistics from multiple batches of production, and determines the actual material weight consumed at each process node during the production of a unit product as the material quota input quantity. For example, for a certain model of metal bracket product, the standard input quantity of aluminum plate for the stamping process node is set to 2.5 kg per piece, and the standard input quantity of welding wire for the welding process node is set to 0.15 kg per piece. The above material quota input quantities are generated by the manufacturing enterprise based on statistics from stable mass production records for no less than three consecutive months, and are updated synchronously after the process parameters are adjusted. After the manufacturing process is read, the corresponding material type is identified based on the material code recorded in the bill of materials. Then, according to the material consumption relationship marked at each process node in the product manufacturing process, the material is allocated to the corresponding process node. For example, if the material code corresponds to aluminum plate raw material, the material is allocated to the stamping process node; if the material code corresponds to welding wire, the material is allocated to the welding process node. If the same material participates in the processing of multiple process nodes consecutively, the allocation is carried out node by node according to the process flow sequence recorded in the production enterprise's process documents. For example, if stainless steel plate continues to enter the bending process after the cutting process, the initial input quantity is recorded at the cutting process node, and the remaining quantity is recorded at the bending process node. After the process allocation is completed, the material quota input quantity corresponding to each process node is calculated based on the corresponding product quantity in the order and the process quota consumption standard. For example, if the order corresponds to 2000 products, and the quota input quantity of a single aluminum plate at the stamping process node is 2.5 kg, then the material quota input quantity corresponding to the stamping process node is determined to be 5000 kg.

[0025] The system reads the process records generated by the manufacturing enterprise during its historical production process. These records detail the processing type and corresponding scrap registration information for each process node. Processing types are defined using the actual process classification method of the manufacturing enterprise, with stamping, CNC cutting, welding, polishing, and painting all treated as independent processing types. Different processing types correspond to different forms of scrap generation. For example, stamping mainly produces metal scrap, CNC cutting mainly produces metal chips and cutting fluid waste, and painting mainly produces paint slag waste. After reading the historical production records, the system extracts the corresponding scrap names from the scrap registration records of the corresponding process nodes and classifies the scrap names by process level based on the process node code. For example, if the stamping process node code is GX01 and the corresponding scrap name is aluminum scrap, then "GX01-Aluminum Scrap" is formed as the theoretical output scrap category; if the welding process node code is GX02 and the corresponding scrap name is welding slag, then "GX02-Welding Slag" is formed as the theoretical output scrap category. By using process code prefixes to distinguish waste names, waste with the same name generated at different process nodes can be independently identified. For example, both cutting and polishing processes may generate metal powder waste, but these are classified into two different theoretical waste categories: "GX03-Metal Powder" and "GX04-Metal Powder," respectively, thus avoiding confusion between different process sources during subsequent waste statistics. For cases where waste names generated at the same process node differ across product models, corresponding theoretical waste categories are established for each product model. For example, if the same painting process generates aluminum paint slag when painting aluminum parts and steel paint slag when painting steel parts, separate records are created for each category. Waste names that appear less frequently than the set statistical frequency in historical production records are not included in the theoretical waste category set.

[0026] Historical production batches are categorized and archived according to product model and process node. Each historical production batch corresponds to a complete production quantity record, material input record, and waste recycling record. For each type of theoretically produced waste, the registered waste recycling weight within the corresponding production batch is read, along with the material quota input for the corresponding process node within the same production batch. For example, if the aluminum plate input for a certain batch's stamping process node is 5000 kg, and the waste weight corresponding to "GX01-Aluminum Scrap" in the waste recycling record is 420 kg, then the waste output rate record for this production batch is the ratio between 420 kg and 5000 kg. Subsequently, the ratio records corresponding to all historical production batches are statistically analyzed in chronological order to form a ratio sequence for the corresponding waste category. To avoid outliers caused by equipment malfunctions, material defects, or temporary process adjustments affecting the overall statistical results, after generating the ratio sequence, outlier records that significantly deviate from the main distribution range are removed. When the ratio of a certain batch is more than twice the average of all historical ratios, the batch is marked as an abnormal production batch and is not included in the average statistics. Finally, the process node, theoretical output waste type and corresponding output rate are written into the process-waste output mapping table.

[0027] In S3, the missing categories of waste materials are extracted.

[0028] Based on the order identifier, all weighing entries generated within the corresponding order delivery cycle are selected from the waste recycling records. The order delivery cycle is determined by the production start date and order completion date registered by the manufacturing company when the order is placed; only waste recycling records within this time range are retained for subsequent statistics. All weighing entries within the order delivery cycle are categorized and aggregated based on waste type. For multiple weighing entries belonging to the same waste type, the corresponding net weight values ​​are summed to form the total actual weight received for that waste type. For example, if there are 5 "GX01-Aluminum Scrap" recycling records within an order delivery cycle, with net weight values ​​of 120 kg, 135 kg, 118 kg, 142 kg, and 125 kg respectively, then the total actual weight received for this waste type is 640 kg after summing these 5 records. During the aggregation process, for cases where the same transport vehicle is weighed repeatedly within a short period, duplicate record verification is performed according to the vehicle number and weighing time interval. In this embodiment, if the same vehicle produces the same type of waste twice within 30 minutes with a weight difference of less than 50 kg, the latter record is marked as a duplicate weighing record and removed, thus preventing the vehicle from being weighed repeatedly and causing an inflated waste count. For ledger entries involving mixed waste, the corresponding weights are assigned based on the classification and sorting records generated by the waste recyclers during collection. For example, if the same vehicle contains both aluminum scraps and copper shavings, they are assigned to their respective waste categories according to the actual weights in the collection and sorting records. After all waste categories are collected, a waste composition distribution corresponding to the current order delivery cycle is formed. The waste composition distribution records the total actual weight received, the cumulative number of weighings, and the number of corresponding transfer vehicles for each waste category.

[0029] The system reads the product production quantity and material input records corresponding to the current order. Then, based on the process nodes, theoretical scrap categories, and corresponding output rates recorded in the process-scrap output mapping table, it calculates the theoretical scrap output weight for each scrap category under normal production conditions for the current order. For example, if the aluminum plate material input for the stamping process node of the current order is 5000 kg, and the output rate for "GX01-Aluminum Scrap" in the process-scrap output mapping table is 9.4%, then the theoretical scrap output weight for this scrap category is calculated to be 470 kg based on the 5000 kg material input and the 9.4% output rate. Considering that differences in equipment status, raw material batches, and manual operation during actual production can lead to normal fluctuations in scrap output, after calculating the theoretical scrap output weight, the lower bound of the theoretical scrap output weight is further determined by combining the scrap fluctuation range in historical production records. In this embodiment, the lower bound of the theoretical scrap output weight is determined by subtracting 15% from the historical average scrap output. For example, if the historical theoretical average output weight of a certain scrap category is 470 kg, then the lower limit of the theoretical scrap output weight is determined to be 399.5 kg. The aforementioned 15% downward adjustment is based on statistics of the manufacturer's continuous and stable production records over the past twelve months, and is set independently for different product types. For example, precision cutting products, due to their higher processing stability, have a downward adjustment of 10%; while stamping products are more affected by batch fluctuations in raw materials, and their downward adjustment is set at 15%. After calculating the lower limit of the theoretical scrap output weight, the actual total weight of the corresponding scrap category in the scrap composition distribution is compared item by item with the lower limit of the theoretical scrap output weight. When the actual total weight is lower than the lower limit of the theoretical scrap output weight, the corresponding scrap category is marked as an absent category. For example, the lower limit of the theoretical scrap output weight for "GX01-aluminum scrap" is 399.5 kg, while the actual total weight received in the scrap composition distribution is only 210 kg; therefore, this scrap category is marked as an absent category.

[0030] In step S4, the process direction for missing product categories that exceed the cross-process waste collaborative fluctuation range is obtained.

[0031] When retrieving the total actual weight of waste materials corresponding to adjacent process nodes within each historical order delivery cycle from a set number of historical order delivery cycles prior to the current order delivery cycle, and constructing a cross-process waste material collaborative fluctuation range, historical order delivery cycles with the same product model and consistent production process are selected from the historical order database based on the product model and production process route corresponding to the current order. To ensure the stability of historical statistical results, the set number of historical order delivery cycles is determined using order records within a continuous and stable production phase of the manufacturing enterprise. In this embodiment, the number of historical order delivery cycles is set to be no less than 20, and it is required that the production equipment, raw material categories, and process parameters of the corresponding historical orders have not undergone significant changes. For example, the stamping die has not been replaced, the welding process has not been modified, and the raw material thickness specifications remain consistent. After completing the historical order screening, the process node information corresponding to the process-waste output mapping table within each historical order delivery cycle is read, and the process adjacency relationship is determined according to the product manufacturing process. For example, the stamping process and the bending process are adjacent process nodes, and the bending process and the welding process are adjacent process nodes. Subsequently, the total actual weight of the theoretical output waste material category corresponding to each adjacent process node in the historical order delivery cycle is read. For example, "GX01-Aluminum Scrap" corresponds to the stamping process node, and "GX02-Aluminum Bending Shavings" corresponds to the bending process node. The total actual weight received for each of these processes in each historical order delivery cycle is then calculated. After this calculation, the ratio between the total actual weight received for each pair of adjacent process nodes is calculated. For example, if the total actual weight received for the stamping process in a historical order delivery cycle is 420 kg, and the total actual weight received for the bending process is 105 kg, the corresponding ratio is recorded as 4.0. Subsequently, the ratio records for all historical order delivery cycles are calculated in chronological order to form a ratio sequence between adjacent process nodes. In this embodiment, when a ratio exceeds twice the average level of all historical ratios, it is considered an abnormal ratio. After removing abnormal values, the minimum and maximum values ​​in the remaining ratio sequence are read and used as the cross-process scrap collaborative fluctuation range between corresponding adjacent process nodes. For example, after statistical analysis, the minimum value of the ratio sequence between the stamping process and the bending process is 3.6 and the maximum value is 4.4. Therefore, 3.6 to 4.4 is determined as the cross-process waste collaborative fluctuation range of this adjacent process node pair.

[0032] The system reads the theoretical output scrap category name corresponding to the missing category and determines the corresponding first process node based on the process code prefix. For example, if the missing category is "GX01-Aluminum Scrap," then the first process node is determined to be the stamping process node corresponding to GX01. Subsequently, the system reads the second process node that has a process flow relationship with the first process node based on the product manufacturing process record. For example, if the subsequent process GX01 stamping process corresponds to the GX02 bending process, then GX02 is taken as the second process node. After determining the adjacent process nodes, the system extracts the actual total weight of the scrap categories corresponding to the first and second process nodes from the scrap composition distribution corresponding to the current order delivery cycle. For example, if the actual total weight of "GX01-Aluminum Scrap" is 210 kg and the actual total weight of "GX02-Aluminum Bending Flakes" is 102 kg in the current order delivery cycle, then the adjacent ratio corresponding to the current order is calculated based on the two actual total weights. In this embodiment, the adjacent ratio is determined by dividing the total actual weight of waste received at the first process node by the total actual weight of waste received at the second process node. In the example above, the adjacent ratio is approximately 2.06. After calculating the adjacent ratio, the adjacent ratio corresponding to the current order is compared with the cross-process waste coordination fluctuation range obtained from historical statistics. For example, if the adjacent ratio corresponding to the current order is 2.06, while the corresponding historical cross-process waste coordination fluctuation range is 3.6 to 4.4, then it is determined that the adjacent ratio corresponding to the current order deviates significantly from the normal process coordination relationship. Since the waste output ratio between adjacent process nodes remains relatively stable under long-term stable production conditions, when the waste corresponding to the first process node decreases significantly, while the waste corresponding to adjacent process nodes remains at a normal level, it indicates that the actual processing behavior of the first process node may not have been completed within the production address registered by the manufacturing enterprise, but rather that some processing tasks have been transferred to external nodes. After completing the fluctuation range comparison, the first process node to which the corresponding missing product category belongs is marked as a suspected outsourcing process position. For example, if the scrap material corresponding to the GX01 stamping process is consistently below the normal collaborative range, then the GX01 stamping process will be marked as a suspected outsourcing process. If multiple suspected outsourcing processes exist in the same order, the sequence of anomalies at the corresponding process nodes will be recorded according to the process flow order, and the magnitude of each process node's deviation from the cross-process scrap material collaborative fluctuation range will be calculated.

[0033] In step S5, the coordinates of non-registered dwelling points whose dwell time exceeds the preset dwelling determination and the corresponding load change sequence are extracted.

[0034] The process node code marked as a suspected outsourcing process is read. For example, if the GX01 stamping process in the current order is identified as a suspected outsourcing process, the theoretical output scrap category name corresponding to the GX01 process node is read from the process-scrap output mapping table. Then, based on the current order identifier and the corresponding theoretical output scrap category name, the corresponding weighing ledger entries are filtered from the scrap recycling records. During the filtering process, only ledger records generated within the order delivery cycle with scrap names consistent with the corresponding theoretical output scrap category are retained. For example, if the current order identifier is DD20260105, only weighing ledger entries with order identifier DD20260105 and scrap name "GX01-Aluminum Scrap" are retrieved. After completing the weighing ledger filtering, the transfer vehicle identifier recorded in each entry is read, and the corresponding vehicle's driving record within the order delivery cycle is retrieved based on the transfer vehicle identifier. The driving record includes the vehicle's location trajectory, trajectory timestamp, and vehicle load record. The vehicle positioning trajectory is continuously collected by a satellite positioning terminal installed on the transport vehicle. In this embodiment, a fixed sampling period of 30 seconds is used, meaning the vehicle uploads its current location coordinates every 30 seconds. The vehicle load record is collected in real time by an on-board weighing device installed at the vehicle's axle. In this embodiment, a 60-second sampling period is used to record the vehicle's total load value. After retrieving the vehicle driving record, the start time corresponding to the current order delivery cycle is read, and the last weighing timestamp in the corresponding weighbridge entry is read. These two timestamps are used as the trajectory interception range. For example, if the start time of the current order delivery cycle is 08:00 on January 5, 2026, and the last weighing timestamp is 18:30 on January 12, 2026, then the vehicle trajectory segment within the corresponding time range is intercepted. After the trajectory interception is completed, the continuous positioning records in the trajectory segment are traversed to identify continuous vehicle dwelling behavior. When the vehicle's positioning coordinate change range remains within a set distance range over multiple consecutive sampling cycles, the vehicle is considered to be in a dwelling state. The distance range is set at 50 meters based on the actual positioning error of the satellite positioning terminal. This means that when the distance between consecutive positioning points does not exceed 50 meters, the vehicle is considered to be in a fixed stopping position. The duration of continuous stopping is then recorded. When the continuous stopping time exceeds a preset stopping threshold, the corresponding stopping position is identified as a valid stopping point. The stopping threshold is set at 20 minutes, specifically based on the historical average loading and unloading time of waste materials at the production enterprise. Trajectory points with a stopping time of less than 20 minutes are not included in the stopping point analysis to avoid misidentification caused by traffic congestion, intersection waiting, or temporary parking. After valid stopping identification is completed, the positioning coordinates of the corresponding stopping position are read and recorded as the stopping point coordinates. For the same vehicle repeatedly entering the same stopping area within a short period, if the interval between two stoppings is less than 30 minutes, the corresponding stopping behaviors are merged into a single continuous stopping event.

[0035] The system retrieves the official production address from the manufacturer's business registration record and converts it into standard geographic coordinates. Since actual production parks have multiple entrances and exits and factory boundaries, the system does not directly use single-point coordinates when matching stop points. Instead, it establishes the boundary of the registered production area based on the actual land area occupied by the manufacturer's factory. The registered production area is established using the center coordinates of the manufacturer's factory as a reference, combined with the actual perimeter wall boundary. When a stop point's coordinates fall within the registered production area, the corresponding stop is considered a normal factory stop and is removed from subsequent analysis. Stop points not falling within the registered production area are retained as non-registered stop point coordinates. For example, if a transport vehicle stops continuously for 45 minutes at a distance of 12 kilometers from the manufacturer's registered factory area, the corresponding stop point is marked as a non-registered stop point. After filtering non-registered stop points, the system retrieves the start and end times of the corresponding stop events and extracts the vehicle's onboard weighing readings corresponding to that time period from the vehicle's driving records. For example, if the corresponding dwell time at a non-registered stop is from 14:20 to 15:05, then the vehicle load records collected continuously within this time range are read. Since vehicles experience continuous weight changes during loading and unloading, all vehicle load readings within the corresponding time period are arranged chronologically to form a load change sequence. For example, if a vehicle enters the stop with a load of 1.2 tons and gradually increases to 3.8 tons during its stay, this forms a continuous load increase sequence. For short-term abnormal fluctuations in vehicle load during the stay, such as minor weight fluctuations caused by personnel getting on and off the vehicle or vibrations from loading / unloading equipment, a valid load change is determined by three consecutive samples showing changes in the same direction. If the cumulative load change after three consecutive samples is less than 100 kg, it is considered an invalid fluctuation and is discarded. If a vehicle engages in both loading and unloading activities while staying at a non-registered stop, the corresponding load change stage is determined based on the direction of the load change.

[0036] In step S6, shared external nodes are identified.

[0037] The system reads the supplier management file corresponding to the manufacturing enterprise. This file records the raw material categories, product models, supply process types, and supply tier information for each supplier. The supply tier is determined based on the actual product processing stage the supplier participates in. For example, suppliers providing rough-processed metal sheets are classified as Level 1 suppliers, while suppliers providing stamped semi-finished products are classified as Level 2 suppliers. After reading the target supplier's supply tier, the system filters suppliers in the same supply tier as the target supplier and those who have a long-term supply history to the same manufacturing enterprise. The filtering criteria for similar suppliers include consistent product categories, consistent corresponding process types, and a continuous supply cycle of at least six months. For example, if the target supplier has a long-term supply history of aluminum stamped structural parts to the manufacturing enterprise, only suppliers that also provide aluminum stamped structural parts are filtered as similar suppliers. After filtering similar suppliers, the system traces back the waste recycling records of corresponding orders within a set historical period, using the current order delivery cycle as the time center. In this embodiment, the historical period is set to 180 days based on the manufacturing enterprise's average order fulfillment cycle, ensuring that orders from different suppliers are at similar production stages. Subsequently, the vehicle travel records from the corresponding waste recycling order are read, and the dwell events corresponding to the coordinates of non-registered dwell points are extracted from the travel records. Dwell events include dwell point coordinates, dwell start time, dwell end time, and the corresponding vehicle load change sequence during the dwell period. For example, if a transport vehicle stays continuously for 50 minutes at a distance of 15 kilometers from the manufacturer's registered factory area, and the vehicle load increases from 1.4 tons to 3.9 tons during the dwell period, the corresponding dwell event and the corresponding load change sequence are recorded.

[0038] The system reads the theoretically generated waste category names corresponding to the identified suspected outsourced work stations. For example, if the waste category name corresponding to the suspected outsourced work station is "GX01-Aluminum Scrap," only the waste recycling records corresponding to "GX01-Aluminum Scrap" are retained for subsequent screening. Then, the system iterates through the load change sequences corresponding to each non-registered stop, performing a unified determination on the load change direction of vehicles during their stop. The load change direction is determined by the difference between the vehicle's load value at the end of the stop and the vehicle's load value at the beginning of the stop. When the ending load value is higher than the beginning load value, it is considered an increasing load direction; when the ending load value is lower than the beginning load value, it is considered a decreasing load direction. Since the waste transfer behavior corresponding to the same outsourced node usually remains relatively stable during long-term operation, only stop events with the same load change direction are retained for the same non-registered stop. For example, if most stop events corresponding to a certain non-registered stop show a continuous increase in vehicle load, only stop events with an increasing load direction are retained, and a small number of abnormal decreasing events are removed. After screening for load change directions, the names of waste categories in the corresponding dwelling events are further verified. When the waste category name matches the theoretically produced waste category name of the suspected outsourced process, the dwelling event is marked as a valid outsourced dwelling event. For example, if a non-registered dwelling point corresponding to "GX01-aluminum scrap" continuously exhibits dwelling behavior with increasing vehicle load and consistent waste category, then the dwelling event is considered to be directly related to the suspected outsourced process. After screening for valid outsourced dwelling events, the total number of dwelling events and the number of valid outsourced dwelling events corresponding to each non-registered dwelling point are counted, and the percentage of valid outsourced dwelling events is calculated. When a non-registered dwelling point has a total of 40 dwelling events, of which 30 are valid outsourced dwelling events, the percentage of valid outsourced dwelling events is 75%. This percentage is then compared with the preset sharing judgment ratio. The preset sharing judgment ratio is determined to be 60% based on the production enterprise's historical normal waste transportation records. When the proportion of valid outsourcing events exceeds 60%, the corresponding non-registered outsourcing points will be marked as shared outsourcing nodes.

[0039] In step S7, the amount of outsourcing for a process is calculated.

[0040] Read the coordinates of the already marked shared outsourcing nodes and filter valid outsourcing dwell events from all dwell events associated with the corresponding shared outsourcing nodes. Valid outsourcing dwell events must simultaneously meet two conditions: the scrap category name must match the theoretical output scrap category name of the suspected outsourcing process position, and the load change direction must be consistent. For example, if the theoretical output scrap category of the suspected outsourcing process position is "GX01-Aluminum Scrap," then only dwell events with the scrap name "GX01-Aluminum Scrap" and continuously increasing vehicle load will be retained. After filtering valid outsourcing dwell events, read the vehicle load change sequence corresponding to each dwell event and extract the initial load value when the vehicle enters the shared outsourcing node and the final load value when leaving the shared outsourcing node. Then, determine the net load increment of the corresponding dwell event by subtracting the initial load value from the final load value. A transport vehicle entered a shared outsourcing node with a load of 1.3 tons and left with a load of 3.7 tons, resulting in a net load increase of 2.4 tons. Since vehicles may briefly unload and reload during their stay, the final stable load value at the end of the stay is used as the final load value when calculating the net load increase. The stable load value is determined based on fluctuations not exceeding 50 kg over three consecutive sampling periods, thus avoiding instantaneous weight deviations caused by vehicle vibrations at the end of loading and unloading. After extracting the net load increase for a single stay event, the net load increases corresponding to all valid outsourcing stay events are categorized and accumulated according to the type of waste material. For example, within the current order delivery cycle, there are 12 valid outsourcing stay events for "GX01-aluminum scrap," and the total weight of the corresponding outsourcing waste after accumulating the net load increases for each event is 18.6 tons.

[0041] The historical output rate of the theoretically produced scrap category corresponding to the suspected outsourced process position is retrieved. For example, if the current suspected outsourced process position is the GX01 stamping process node, the historical output rate of the corresponding theoretically produced scrap category "GX01-aluminum scrap" is 9.4%. Then, the total weight of outsourced scrap accumulated from the corresponding shared outsourced nodes is retrieved, and the processing volume of the corresponding process is deduced based on the process scrap output relationship. The outsourced quantity of the process is determined by dividing the total weight of outsourced scrap by the corresponding scrap output rate. For example, if the total weight of "GX01-aluminum scrap" outsourced scrap accumulated from the shared outsourced nodes is 18.6 tons, and the corresponding historical output rate is 9.4%, then the calculated outsourced quantity for the corresponding process is approximately 197.9 tons. Since there are normal fluctuations in scrap output rates between different production batches, after calculating the outsourced quantity for the process, a reasonableness check is further performed based on the fluctuation range of the historical output rate of the corresponding process node. For example, if the historical output rate of process node GX01 remains stable between 8.8% and 10.1%, the outsourcing volume range for the process will be calculated synchronously based on this fluctuation range. When the difference between the outsourcing volume calculated from the lowest and highest historical output rates exceeds 20%, the corresponding shared outsourcing node will be marked as an abnormal fluctuation node, and the original load records for the corresponding time period will be retrieved for re-verification. After completing the outsourcing volume calculation, the location coordinates, outsourcing volume, and identifiers of similar suppliers associated with the shared outsourcing node will be output. The similar supplier identifiers will be recorded using supplier registration numbers. For example, if a shared outsourcing node is associated with three aluminum stamping parts suppliers at the same material supply level, the registration numbers of the three suppliers will be output synchronously in the results. For cases where the same shared outsourcing node appears repeatedly in multiple order delivery cycles, the cumulative number of occurrences and cumulative outsourcing volume of the corresponding node will be further recorded. When the same shared outsourcing node appears more than 15 times in a total of 6 consecutive months, the node is marked as a long-term shared outsourcing node, and the corresponding cumulative outsourcing statistics are output synchronously in the results.

[0042] Example 2 The difference between Embodiment 2 and Embodiment 1 is that this embodiment introduces a big data-based intelligent management system for the industrial chain.

[0043] Figure 2 A schematic diagram of a big data-based intelligent supply chain management system is provided. The big data-based intelligent supply chain management system includes: The data acquisition module is used to acquire the bill of materials and the weighing ledger in the waste recycling records corresponding to the order delivery cycle between the production enterprise and the target supplier, and associate the order identifier, material code, transfer vehicle identifier and transfer vehicle driving record to establish the correspondence between orders, waste and transfer vehicles. The process waste analysis module is used to establish a mapping relationship between the output rate of each process node and the theoretical output waste category based on the material consumption relationship and historical production records of each process node in the product manufacturing process, form a process-waste output mapping table, and identify missing categories that are below the lower limit of the theoretical waste output weight. The outsourcing identification module is used to calculate the cross-process waste coordination fluctuation range between adjacent process nodes based on historical production records, and to perform coordination fluctuation analysis on the total actual weight of waste received between the process node corresponding to the missing product category and adjacent process nodes to identify suspected outsourcing process positions. The shared outsourcing traceability module is used to extract the load change sequence corresponding to non-registered depots, count the proportion of effective outsourcing depot events of transfer vehicles from different suppliers of the same type at non-registered depots, identify shared outsourcing nodes, and calculate the outsourcing quantity of processes in combination with the process-waste output mapping table.

[0044] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0045] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0046] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0047] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0048] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0049] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0050] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0051] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0052] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A big data-based intelligent management method for the industrial chain, characterized in that, Includes the following steps: S1. Obtain the bill of materials between the manufacturer and the target supplier, as well as the weighing ledger in the waste recycling records, within the order delivery cycle; S2. Using the bill of materials as the decomposition object, break down the material input of each process node along the product manufacturing process. Based on historical production records, establish a mapping relationship between the process node and the output rate corresponding to the theoretical output waste category, and form a process-waste output mapping table. S3. Aggregate the weighing ledger by waste category into the waste composition distribution of the order delivery cycle, and align it with the process-waste output mapping table to extract the missing waste categories. S4. Based on historical production records, calculate the collaborative fluctuation range corresponding to the waste output rate of each adjacent process node, and obtain the process pointers of the missing categories of waste output rate that exceed the cross-process collaborative fluctuation range, as suspected outsourcing process positions. S5. Using the suspected outsourced work station as an index, extract the coordinates of non-registered dwelling points and the corresponding load change sequence from the transportation vehicle driving records of waste recycling records within the current order delivery cycle; S6. Retrieve the waste recycling records of similar suppliers of the target supplier during their order delivery cycle, count the proportion of valid outsourcing events corresponding to non-registered outsourcing point coordinates, and identify shared outsourcing nodes. S7. Based on the net increase in load of transfer vehicles in the shared outsourcing nodes, calculate the outsourcing quantity of the process in combination with the process-waste output mapping table, and output the coordinates of the shared outsourcing nodes, the outsourcing quantity of the process, and the set of related similar suppliers.

2. The intelligent supply chain management method based on big data according to claim 1, characterized in that, In step S1, obtaining the bill of materials between the manufacturer and the target supplier, as well as the weighing ledger in the waste recycling record, during the order delivery cycle specifically includes: The bill of materials is generated within the delivery cycle after the manufacturing enterprise places an order with the target supplier. The list entries record the material codes, order identifiers and corresponding material quantities that constitute the product production. Retrieve the weighing ledger entries within the order delivery cycle corresponding to the current order. The weighing ledger entries record the waste material name, order identifier, and corresponding transfer vehicle identifier. The driving records of the corresponding transfer vehicle during the order delivery period are retrieved based on the vehicle identification. The driving records include the vehicle location trajectory, timestamp, and vehicle load record.

3. The intelligent supply chain management method based on big data according to claim 1, characterized in that, In step S2, the process-waste output mapping table is formed specifically including: Based on the material consumption relationships marked at each process node in the product manufacturing process of the manufacturing enterprise, the materials in the bill of materials are allocated to the corresponding process nodes according to the material code and the material quota input quantity; Based on the processing type of each process node, the names of the waste materials generated by the corresponding process node are extracted from the product's historical production records. After adding a unique process code prefix to the waste material names, they are used as the theoretical output waste material categories for the corresponding process node. In the product's historical production records, the ratio sequence of waste weight to material quota input for each waste category in each production batch is statistically analyzed. The mean of the ratio sequence is extracted as the output rate of the corresponding waste category, and a process-waste output mapping table is established.

4. The intelligent supply chain management method based on big data according to claim 1, characterized in that, In S3, the missing categories of waste materials to be extracted specifically include: Read the waste categories recorded in each entry of the weighing ledger corresponding to the current order identifier, and classify and collect all weighing ledger entries within the order delivery cycle based on the waste categories to establish the waste composition distribution. Iterate through the total actual weight of each waste category in the waste composition distribution, and combine it with the output rate of the corresponding waste category in the process-waste output mapping table to calculate the theoretical waste output weight. Mark the waste categories whose total actual weight is lower than the lower bound of the theoretical waste output weight as missing categories.

5. The intelligent supply chain management method based on big data according to claim 1, characterized in that, In step S4, obtaining the process pointer for missing product categories that exceed the cross-process waste collaborative fluctuation range specifically includes: From the historical order delivery cycles set before the current order delivery cycle, retrieve the total actual weight of each type of waste produced by adjacent process nodes in the process-waste output mapping table within each historical delivery cycle. Calculate the ratio sequence of the total actual weight of the two types of waste for each pair of adjacent process nodes. The minimum and maximum values ​​of the ratio sequence constitute the cross-process waste collaborative fluctuation range of the adjacent process node pair. Based on the process-waste output mapping table, read the first process node corresponding to the missing product category and the second process node that has a process adjacency relationship with the first process node, extract the actual total weight of waste products of the first process node and the second process node in the current order delivery cycle, and calculate the adjacency ratio. If the adjacent ratios do not fall within the cross-process waste collaborative fluctuation range corresponding to the first process node and the second process node, then the first process node corresponding to the missing product category will be marked as a suspected outsourcing process position.

6. The intelligent supply chain management method based on big data according to claim 1, characterized in that, In step S5, extracting the coordinates of non-registered dwell points whose dwell time exceeds the preset dwell time determination and the corresponding load change sequence specifically includes: Read the theoretical waste category name corresponding to the suspected outsourced process position in the process-waste output mapping table, retrieve the ledger entry from the weighing ledger by category name and order identifier, and obtain the driving record of the transfer vehicle; The trajectory segment from the start of the current order delivery cycle to the last weighing timestamp in the driving record of the transfer vehicle is extracted. The stopping positions where the dwell time exceeds the preset dwell judgment threshold are extracted from the trajectory segment, and the location coordinates of the stopping positions are used as the dwell point coordinates. The coordinates of the stopping points that are consistent with the production address registered by the manufacturer are removed, the coordinates of the non-registered stopping points are retained, and the on-board weighing readings that are consistent with the time period of the non-registered stopping points are extracted from the driving records to form a load change sequence.

7. The intelligent supply chain management method based on big data according to claim 1, characterized in that, In step S6, identifying shared external nodes specifically includes: Screen similar suppliers at the same material supply level as the target supplier, retrieve the waste recycling records corresponding to the orders placed by the production enterprise to these similar suppliers within a set historical period, and extract the stationing events and load change sequences of the transfer vehicles at the coordinates of non-registered stationing points. Filter out the dwell events in the load change sequence where the direction of change is consistent and the waste category name is consistent with the theoretical output waste category name of the suspected outsourced process station, and record such dwell events as valid outsourced dwell events. The proportion of valid outsourcing events among the total events corresponding to non-registered outsourcing point coordinates is counted. If the proportion exceeds the preset sharing judgment ratio, the corresponding non-registered outsourcing point coordinates are marked as shared outsourcing nodes.

8. The intelligent supply chain management method based on big data according to claim 1, characterized in that, In step S7, calculating the outsourcing quantity of the process specifically includes: Extract valid outsourcing stay events from shared outsourcing nodes, and sum the net increase in the load capacity of the transfer vehicle corresponding to each valid outsourcing stay event according to the waste type to obtain the total weight of outsourced waste. Read the output rate of the suspected outsourced process position from the process-waste output mapping table, divide the total weight of outsourced waste by the output rate to obtain the outsourced quantity of the process, and output the location coordinates of the shared outsourced node, the outsourced quantity of the process, and the identifier of the associated similar supplier corresponding to the shared outsourced node.

9. A big data-based intelligent supply chain management system, used to implement the big data-based intelligent supply chain management method according to any one of claims 1-8, characterized in that, include: The data acquisition module is used to acquire the bill of materials and the weighing ledger in the waste recycling records corresponding to the order delivery cycle between the production enterprise and the target supplier, and associate the order identifier, material code, transfer vehicle identifier and transfer vehicle driving record to establish the correspondence between orders, waste and transfer vehicles. The process waste analysis module is used to establish a mapping relationship between the output rate of each process node and the theoretical output waste category based on the material consumption relationship and historical production records of each process node in the product manufacturing process, form a process-waste output mapping table, and identify missing categories that are below the lower limit of the theoretical waste output weight. The outsourcing identification module is used to calculate the cross-process waste coordination fluctuation range between adjacent process nodes based on historical production records, and to perform coordination fluctuation analysis on the total actual weight of waste received between the process node corresponding to the missing product category and adjacent process nodes to identify suspected outsourcing process positions. The shared outsourcing traceability module is used to extract the load change sequence corresponding to non-registered depots, count the proportion of effective outsourcing depot events of transfer vehicles from different suppliers of the same type at non-registered depots, identify shared outsourcing nodes, and calculate the outsourcing quantity of processes in combination with the process-waste output mapping table.

Images