Data acquisition and treatment method for new energy automobile chip industry

Abstract

The invention discloses a new energy automobile chip industry data acquisition and treatment method. The method comprises the following steps: acquiring multi-source heterogeneous data of an EDA tool chain, a wafer factory MES system, a supply chain SaaS platform and a market transaction database in real time through a distributed data pipeline; constructing a chip industry knowledge graph including enterprise entities, product models, supply chain nodes, substitution relationships and risk conduction paths; a dynamic weight model is adopted to clean and aggregate the data, and the dynamic weight model combines the price sensitivity and the delivery time emergency degree and adjusts an adaptive coefficient in real time according to the market supply-demand ratio; classifying and storing the treated data into a time sequence database, a graph database and a column database according to scenes; and triggering supply chain risk early warning based on the probability model of multi-factor fusion and generating a three-level substitution chip scheme. According to the invention, the problems of incomplete data acquisition, rigid management, lagging risk early warning and lack of alternative schemes in the prior art are solved, and the stability of a new energy automobile chip supply chain is ensured.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method for collecting and managing data for the new energy vehicle chip industry. Background Technology

[0002] In the global industrial landscape, the new energy vehicle industry is booming, becoming a key force leading the future transformation of transportation. New energy vehicles, with their advantages of environmental friendliness and high efficiency, are seeing their market share continue to climb, and the industry scale is constantly expanding. The new energy vehicle chip industry, as a core support of the new energy vehicle industry, involves the entire industrial chain from design and production to sales and application. Each link generates a large amount of data, which is characterized by being multi-source, heterogeneous, and massive.

[0003] Currently, while some technologies exist for data processing in the new energy vehicle chip industry, numerous problems remain. In the data acquisition stage, some companies employ traditional methods, extracting data from various business systems, resulting in incomplete data collection and poor real-time performance. In the data governance stage, although some companies have established data warehouses and data lakes, their capabilities in data quality control, consistency maintenance, and value extraction are insufficient. Regarding risk warning and alternative solutions, existing technologies struggle to accurately predict supply chain risks and lack effective chip alternatives. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art and proposes a method for data collection and management in the new energy vehicle chip industry.

[0005] In a first aspect, embodiments of the present invention provide a method for data collection and management in the new energy vehicle chip industry, including:

[0006] Real-time data collection of the new energy vehicle chip industry through distributed data pipelines; specifically including: obtaining chip design parameters, including power consumption, computing power, and temperature drift coefficient, from the EDA toolchain; extracting capacity utilization, yield rate, and work-in-process status from the wafer fab MES system; synchronizing order delivery dates, inventory depth, and logistics node information from the supply chain SaaS platform; and capturing price fluctuation time-series data and corporate public opinion events from the market transaction database.

[0007] Based on the aforementioned chip industry data, a chip industry knowledge graph is constructed, linking enterprise entities, chip product models, and supply chain nodes into a topological network. Enterprise entities include design companies, wafer fabs, packaging and testing plants, and automotive-grade certification bodies. Chip product models are associated with multi-level BOM lists and substitution relationship matrices. Supply chain nodes are marked with geographical coordinates, transportation routes, and capacity elasticity coefficients.

[0008] Dynamic weight governance is performed on the knowledge graph to obtain a weight-governed dataset. Specifically, this includes: loading a dynamic weight model for each data record in the knowledge graph, wherein the model calculation rule is: weight value W = α × price sensitivity + β × delivery urgency, where α and β are adaptive coefficients that are adjusted in real time according to the market supply and demand ratio; cleaning conflicting data based on the weight value W, retaining data with a weight value W greater than a threshold T, and aggregating cross-source data to generate the governed dataset.

[0009] The processed datasets are categorized and stored according to industry scenarios. Specifically, the time-series database stores price fluctuations and inventory dynamics data; the graph database stores supply chain topology relationships; and the columnar database stores chip performance parameter sets.

[0010] Based on stored data, supply chain risk warnings are triggered, and the probability of delivery delays and alternative chip solutions are output.

[0011] Furthermore, the entity relationship chain of the aforementioned chip industry knowledge graph adopts a four-layer structure: where:

[0012] The first layer uses the design company ID as the root node, connecting the set of chip models it is bound to, forming a master-slave mapping relationship between the design source and the product model;

[0013] The second layer maps each chip model to the corresponding wafer manufacturing process node and the required mask ID, forming a vertical extension of manufacturing parameters;

[0014] The third layer uses the wafer process node as an intermediary entity, and links it downwards to the packaging and testing plant number and the key soldering material code, realizing the extension of information from the process to the packaging and testing stage;

[0015] The fourth layer starts with the packaging and testing plant and connects to the automotive-grade certification certificate number and its expiration date, supporting traceability management of compliance and certification timeliness.

[0016] Furthermore, the conditions for establishing chip model substitution relationships include:

[0017] The power consumption deviation between the two chip models satisfies PA−PB / PA≤0.1, where PA and PB are the rated power consumption of chip A and chip B, respectively, and their temperature drift coefficients are consistent. The two chip models are consistent in the automotive-grade certification path, satisfying that they share the same certification certificate number or the same certification body. After satisfying the above conditions, a substitute edge is established between the two chip models in the knowledge graph. This edge has directional and conditional constraint attributes, and the edge weight participates in the calculation of the dynamic weight model.

[0018] Furthermore, based on the upstream and downstream traversal mechanism of multi-level entity relationships in the knowledge graph, dynamic annotation of supply chain risk transmission paths is achieved. Specific methods include:

[0019] When a production interruption, equipment failure, or geopolitical event is detected at a wafer fab, a three-level recursive analysis of the relevant nodes in the knowledge graph is triggered: The first level analyzes the key chip models whose supply capacity accounts for more than 30% of the total capacity of a certain chip model at the wafer fab; the second level analyzes the position of the affected chip models in the OEM's BOM list and the corresponding order quantity; the third level analyzes the chip models that meet the substitutable conditions and their current inventory position at the packaging and testing plant.

[0020] Assign an impact strength value to each risk transmission path. The calculation formula is: Where E represents the event level, which is classified according to the number of days the wafer fab is shut down: E=1 for less than 3 days, E=2 for 3 to 7 days, and E=3 for more than 7 days, reflecting the basic severity of the event; Rs represents the inventory fulfillment rate of replacement chips, calculated as follows: Paths with a response intensity higher than a preset threshold will be dynamically marked in the knowledge graph with a preset color.

[0021] Furthermore, the calculation process for the price sensitivity includes:

[0022] Using a preset time window as the monitoring window, time-series data of chip prices are extracted from the market transaction database. The price curve is fitted using cubic spline interpolation, and the price change per unit time is calculated. Where ΔP represents the price change of the fitted price curve between the start and end points of the window, in yuan; Δt represents the time span of the monitoring window, which is fixed at 5 minutes.

[0023] Extract trading volume data within the same monitoring window, calculate the standard deviation of trading volume σV, and divide it by the mean of trading volume. To obtain the trading volume volatility ,

[0024] Set a public opinion event compensation factor γ. When the monitoring window detects that the text contains preset risky words, γ=1.5; otherwise, γ=1. Price sensitivity. The calculation formula is:

[0025] .

[0026] Furthermore, the calculation process for delivery urgency includes:

[0027] Extracting the standard delivery date corresponding to the target chip model from the knowledge graph. and the current or projected capacity utilization rate of the wafer fabs associated with the chips. ;

[0028] Dynamic inventory depth is calculated based on inventory information obtained from the supply chain platform and recent consumption data. The formula is ,in, Current inventory level This refers to the average daily shipments or consumption over the past 7 days; if Activate the inventory warning flag on the same day;

[0029] urgency of delivery The calculation is performed using the following formula:

[0030]

[0031] Where the denominator This indicates the combined effect of inventory protection capacity and production capacity supply capacity. Normalize to the interval [0.1, 10].

[0032] Furthermore, cross-source data is aggregated to generate a post-governance dataset. Specific methods include:

[0033] Inventory information for the same chip model is aggregated, and the original inventory value set is obtained from k independent data sources. Based on dynamic weight set Filter and retain The data source and the effective inventory calculation formula are as follows:

[0034]

[0035] Consistency verification is performed on wafer fab capacity data. The three data sources with the highest weight values ​​are selected to form a candidate set. The sample mean μ and standard deviation σ are calculated. Data points with relative mean deviations exceeding twice the standard deviation are removed based on the Laida criterion. The remaining data are filled into the knowledge graph using the arithmetic mean.

[0036] Furthermore, the time-series database is stored using an optimized strategy of partitioned compression and multi-granularity aggregation, specifically including:

[0037] The system partitions data based on the chip model field, and maintains a three-tiered data retention mechanism within each partition: High-frequency raw data is collected at 5-minute intervals, retaining the most recent 7 days; data is aggregated in 1-hour windows, calculating the mean, maximum, and minimum values, and retaining them for 30 days; price volatility is calculated daily.

[0038]

[0039] in, For volatility, The price of each transaction within the day. The system records this indicator as the daily average price, with a 1-day granularity and retains it for 2 years to support cross-cycle risk modeling and supply strategy evaluation.

[0040] A two-tiered indexing mechanism is constructed: the primary index uses a combination key of chip model and timestamp; the secondary index sets high-risk marking logic based on price change gradients, whereby any record meets the criteria... hour, The price at the previous moment is automatically labeled as "high risk".

[0041] Furthermore, the graph database implements relation indexing by constructing a bidirectional adjacency matrix, specifically including:

[0042] The adjacency matrix is ​​mapped along the path of "enterprise → product model → upstream raw material supplier" to build a vertical traceability path for the manufacturing chain; the column direction is mapped along the path of "enterprise → downstream customer → vehicle manufacturer" to track the distribution of chip products in end applications and order mapping relationships.

[0043] Introducing relational weight attributes, supply weight ,in, This indicates the percentage of the annual procurement amount for the corresponding raw materials or components. Indicates logistics dependence; substitution weight , For certification compatibility scoring, the score is based on whether the chip has passed the same certification body or holds a valid automotive-grade certificate. Mp represents the performance matching degree, which quantifies the similarity of the chips in terms of power consumption, temperature drift, and interface parameters.

[0044] Furthermore, based on stored data, supply chain risk warnings are triggered, outputting delivery delay probabilities and alternative chip solutions. Specific methods include:

[0045] Calculate the comprehensive risk factor: ;in, This is an inventory warning flag, with a value of 0 or 1, indicating whether the inventory is below the safety threshold. This is a capacity overload indicator, with a value ranging from 0 to 1, representing the degree of production line load. This is a logistics risk value, reflecting the stability of the route and the level of external risk.

[0046] Introducing the Sigmoid function to calculate the probability of delivery delay: , where K is a positive real parameter that adjusts the steepness of the curve;

[0047] Perform three-level alternative chip matching: Level 1 matching filters chip models whose power consumption and operating voltage deviation are within the preset range; Level 2 matching prioritizes models that are fully pin-to-pin compatible, and if none are available, an adapter board solution is allowed with the adapter cost limited to the preset price; Level 3 matching prioritizes recommending chips with the same certification certificate number.

[0048] This invention discloses a comprehensive heterogeneous data collection and dynamic governance method for the new energy vehicle chip industry. It collects multi-source heterogeneous data in real time from EDA toolchains, wafer fab MES systems, supply chain SaaS platforms, and market transaction databases via a distributed data pipeline. It constructs a chip industry knowledge graph including enterprise entities, product models, supply chain nodes and substitution relationships, and risk transmission paths. A dynamic weighting model is used to clean and aggregate the data, combining price sensitivity and delivery urgency, and adjusting adaptive coefficients in real time based on the market supply-demand ratio. The governed data is categorized and stored in time-series, graph, and columnar databases according to scenarios. A multi-factor fusion probabilistic model triggers supply chain risk warnings and generates three-tiered alternative chip solutions. This invention solves the problems of incomplete data collection, rigid governance, delayed risk warnings, and lack of alternative solutions in existing technologies, ensuring the stability of the new energy vehicle chip supply chain. Attached Figure Description

[0049] Figure 1 A flowchart illustrating a method for data collection and management in the new energy vehicle chip industry, provided as an embodiment of the present invention;

[0050] Figure 2 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0051] To enable those skilled in the art to better understand the technical solutions of the present invention, exemplary embodiments of the present invention are described below in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0052] Where there is no conflict, the various embodiments of the present invention and the features thereof may be combined with each other.

[0053] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0054] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Terms such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0055] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant art and the invention, and will not be interpreted as having an idealized or overly formal meaning unless expressly so defined herein.

[0056] In the technical solution of this invention, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information all comply with relevant laws and regulations and do not violate public order and good morals. The use of user data in this technical solution follows relevant national laws and regulations (e.g., the "Information Security Technology - Personal Information Security Specification"). For example: appropriate measures are taken for personal information access control; restrictions are imposed on the display of personal information; the purpose of using personal information does not exceed the scope of direct or reasonable association; and explicit identity targeting is eliminated when using personal information to avoid precisely locating a specific individual.

[0057] Existing technologies disclose a data quality detection method and apparatus for multidimensional heterogeneous data. This method standardizes resource data received from different data sources and heterogeneous data located in different network domains and business domains. It further identifies the acquired resource data by calling different processing interfaces to achieve data completion, and then performs quality detection on the completed data to determine whether to store it or issue an alarm. However, this method only achieves general data standardization and quality detection (such as null value filling and format unification), and does not model specific data such as design parameters (power consumption / temperature drift), supply chain levels (wafer fab - packaging and testing plant - automotive certification), and substitution relationships in the chip industry. Its dynamic governance logic is simplistic, unable to respond to real-time fluctuations in the chip market, and lacks quantitative analysis of supply chain risk transmission.

[0058] Existing technologies also disclose multi-source, multi-dimensional data acquisition, aggregation, and management methods and equipment for areas with a high proportion of new energy hydropower. This method collects multiple types of data through data acquisition and monitoring control systems such as SCADA within the target area. It integrates wind farm data, photovoltaic power station data, hydropower station data, and grid-connected data using a neural network-based data preprocessing method with unified standards and specifications. Secondly, it establishes a real-time data cloud platform, indexes metadata through data cubes, forms a unified architecture storage model of "phenomenon-data-data source," and develops corresponding equipment. However, its knowledge graph lacks fine-grained product-level associations, lacks alternative analysis capabilities, lacks market factor-driven dynamic governance, and has a static weight model.

[0059] To address at least one of the technical problems existing in the aforementioned related technologies, this embodiment provides a method for data collection and management in the new energy vehicle chip industry, such as... Figure 1 ,include:

[0060] S100. Real-time data collection of the new energy vehicle chip industry through distributed data pipelines; specifically including: obtaining chip design parameters, including power consumption, computing power, and temperature drift coefficient, from the EDA toolchain; extracting capacity utilization, yield, and work-in-process status from the wafer fab MES system; synchronizing order delivery dates, inventory depth, and logistics node information from the supply chain SaaS platform; and capturing price fluctuation time-series data and corporate public opinion events from the market transaction database.

[0061] The distributed data pipeline adopts a distributed architecture, which can connect to multiple data sources simultaneously to achieve parallel data acquisition, greatly improve data acquisition efficiency, ensure data real-time performance, and avoid the problems of data lag and incomplete coverage in the traditional single data source acquisition method.

[0062] S200. Construct a chip industry knowledge graph based on the aforementioned chip industry data, linking enterprise entities, chip product models, and supply chain nodes into a topological network. Enterprise entities include design companies, wafer fabs, packaging and testing plants, and automotive-grade certification bodies. Chip product models are associated with multi-level BOM lists and substitution relationship matrices. Supply chain nodes are marked with geographical coordinates, transportation routes, and capacity elasticity coefficients. Specifically, enterprise entities cover key participants in the chip industry, such as design companies, wafer fabs, packaging and testing plants, and automotive-grade certification bodies, clearly defining the role and positioning of each enterprise in the industry chain. Product models are associated with multi-level BOM lists and substitution relationship matrices for each chip product model. The multi-level BOM lists clearly display the components and hierarchical relationships of the product, and the substitution relationship matrix records information on replaceable chip models. Supply chain nodes are marked with geographical coordinates, transportation routes, and capacity elasticity coefficients. Geographical coordinates facilitate node location, transportation routes provide a basis for logistics planning, and capacity elasticity coefficients reflect the node's ability to adjust capacity.

[0063] In this embodiment, the entity relationship chain of the chip industry knowledge graph is constructed using a four-layer structure: wherein:

[0064] The first layer uses the design company ID as the root node, connecting the set of chip models it is bound to, forming a master-slave mapping relationship between the design source and the product model;

[0065] The second layer maps each chip model to the corresponding wafer manufacturing process node and the required mask ID, forming a vertical extension of manufacturing parameters;

[0066] The third layer uses the wafer process node as an intermediary entity, and links it downwards to the packaging and testing plant number and the key soldering material code, realizing the extension of information from the process to the packaging and testing stage;

[0067] The fourth layer starts with the packaging and testing plant and connects to the automotive-grade certification certificate number and its expiration date, supporting traceability management of compliance and certification timeliness.

[0068] In this embodiment, the conditions for establishing a chip model substitution relationship include: the power consumption deviation between the two chip models satisfies PA−PB / PA≤0.1, where PA and PB are the rated power consumption of chip A and chip B, respectively, and their temperature drift coefficients are consistent; the two chip models are consistent on the automotive-grade certification path, satisfying the requirement of sharing the same certification certificate number or the same certification body; after satisfying the above conditions, a substitutable edge is established between the two chip models in the knowledge graph. This edge has directional and conditional constraint attributes, and its edge weight participates in the calculation of the dynamic weight model. Specifically, the "substitutable" edge has directional and conditional constraint attributes, and its edge weight participates in the calculation of the dynamic weight model in step S300, providing a basis for the subsequent generation of substitution schemes.

[0069] Based on the upstream and downstream traversal mechanism of multi-level entity relationships in knowledge graphs, dynamic annotation of supply chain risk transmission paths is achieved. Specific methods include:

[0070] When a production interruption, equipment failure, or geopolitical event is detected at a wafer fab, a three-level recursive analysis of the relevant nodes in the knowledge graph is triggered: The first level analyzes the key chip models whose supply capacity accounts for more than 30% of the total capacity of a certain chip model at the wafer fab; the second level analyzes the position of the affected chip models in the OEM's BOM list and the corresponding order quantity; the third level analyzes the chip models that meet the substitutable conditions and their current inventory position at the packaging and testing plant.

[0071] Assign an impact strength value to each risk transmission path. The calculation formula is: Where E represents the event level, which is classified according to the number of days the wafer fab is shut down: E=1 for less than 3 days, E=2 for 3 to 7 days, and E=3 for more than 7 days, reflecting the basic severity of the event; Rs represents the inventory fulfillment rate of replacement chips, calculated as follows: The system dynamically marks paths with impact intensity exceeding a preset threshold in the knowledge graph using a preset color. The numerator is taken from the total inventory of packaging and testing plant nodes associated with alternative chips in the knowledge graph, and the denominator is the aggregated value of OEM order demand in secondary associations. When Rs→1, it indicates that the current market has complete substitution capabilities, and the path risk can be effectively hedged; when Rs→0, it indicates a lack of alternative chips, no risk mitigation mechanism, and the path impact value is close to the upper limit of the event level. Finally, the system dynamically marks paths with impact intensity exceeding the preset threshold in red in the graph and pushes this information as input to the risk warning module in step S500, realizing visualized tracking and intelligent response to the spread of disruptions in the new energy vehicle chip industry chain, providing enterprises with accurate substitution scheduling basis and order correction strategies.

[0072] S300. Perform dynamic weight governance on the knowledge graph to obtain a weight-governed dataset; specifically, this includes: loading a dynamic weight model for each data record in the knowledge graph, wherein the model calculation rule is: weight value W = α × price sensitivity + β × delivery urgency, where α and β are adaptive coefficients that are adjusted in real time according to the market supply and demand ratio; cleaning conflicting data based on the weight value W, retaining data with a weight value W greater than the threshold T, and aggregating cross-source data to generate the governed dataset;

[0073] In this embodiment, the calculation process for price sensitivity includes:

[0074] Using a preset time window as the monitoring window, time-series data of chip prices are extracted from the market transaction database. The price curve is fitted using cubic spline interpolation, and the price change per unit time is calculated. Where ΔP represents the price change of the fitted price curve between the start and end points of the window, in yuan; Δt represents the time span of the monitoring window, which is fixed at 5 minutes.

[0075] Extract trading volume data within the same monitoring window and calculate the standard deviation σ of trading volume. V Divide by the average trading volume To obtain the trading volume volatility , where σ V This represents the standard deviation of the trading volume within the window, reflecting the instability of trading activity. This is the arithmetic average of the transaction volume within this window.

[0076] Set a public opinion event compensation factor γ. When the monitoring window detects that the text contains preset risky words, γ=1.5; otherwise, γ=1. Price sensitivity. The calculation formula is:

[0077]

[0078] in, For volatility, The price of each transaction within the day. The system records this indicator as the daily average price at a 1-day granularity and retains it for 2 years to support cross-cycle risk modeling and supply strategy evaluation. To control the impact of abnormal fluctuations on model weights, The value is restricted to the range [0.1, 5.0], and any value exceeding this range is truncated at the boundary.

[0079] This price sensitivity, as part of the dynamic weighting model, participates in the calculation of the following formula:

[0080]

[0081] Where W is the comprehensive weight value of the data record; α and β are coefficients that are adaptively adjusted according to the supply-demand ratio; This indicates the urgency of delivery. This mechanism ensures that records of highly volatile prices and active markets receive higher retention priority in subsequent data governance, enhancing the system's dynamic awareness of risks in the chip supply chain.

[0082] In this embodiment, the calculation of delivery urgency is based on the fusion processing of knowledge graph and multi-source heterogeneous data. First, the standard delivery date T corresponding to the target chip model is extracted from the knowledge graph. std The unit is days. This parameter is derived from historical procurement cycle statistics and reflects the time from order placement to delivery under normal supply conditions. It also extracts the current or projected capacity utilization rate U of the wafer fab associated with the chip. cap This value, ranging from 0 to 100, represents the current capacity utilization level. A higher value indicates that the production line is closer to saturation and the supply capacity is tighter. The calculation process for delivery urgency includes:

[0083] Extracting the standard delivery date corresponding to the target chip model from the knowledge graph. and the current or projected capacity utilization rate of the wafer fabs associated with the chips. ;

[0084] Dynamic inventory depth is calculated based on inventory information obtained from the supply chain platform and recent consumption data. The formula is ,in, This represents the current inventory level, in units of pieces. This represents the average daily shipments or consumption over the past 7 days. On that day, the system activated an inventory warning flag, indicating a potential risk of supply disruption. Simultaneously, based on transportation routes and node information marked in the knowledge graph, it identified whether the route traversed earthquake zones or areas of geopolitical conflict, assigning a transportation risk coefficient η, with a value of 0.7 or 1.0, to quantify the impact of logistics route stability on delivery risk.

[0085] urgency of delivery The calculation is performed using the following formula:

[0086]

[0087] Where the denominator This indicates the combined effect of inventory protection capacity and production capacity supply capacity. Normalize to the interval [0.1, 10].

[0088] The urgency of delivery is the second term in the weighting model, participating in the calculation of the overall weight:

[0089]

[0090] in The price sensitivity calculated in the previous step is represented by α and β, which are dynamically adjusted weighting coefficients. By introducing three factors—inventory pressure, capacity constraints, and logistics risks—this delivery urgency index enhances the system's responsiveness to short-term supply disruptions and structural shortage risks, enabling precise management of fluctuations in the chip supply chain.

[0091] In this embodiment, the method for aggregating cross-source data to generate a post-governance dataset includes:

[0092] Inventory information for the same chip model is aggregated, and the original inventory value set is obtained from k independent data sources. Based on dynamic weight set Filter and retain The data source and the effective inventory calculation formula are as follows:

[0093]

[0094] Consistency verification is performed on wafer fab capacity data. The three data sources with the highest weight values ​​are selected to form a candidate set. The sample mean μ and standard deviation σ are calculated. Data points with relative mean deviations exceeding twice the standard deviation are removed based on the Laida criterion. The remaining data are filled into the knowledge graph using the arithmetic mean.

[0095] Specifically, This represents the aggregated effective inventory value, possessing cross-source credibility weighting capabilities to enhance the accuracy and robustness of the inventory depth indicator. Subsequently, consistency checks are performed on wafer fab capacity data from multiple sources. If conflicts are found in the capacity utilization data of key nodes, the system selects the three data sources with the highest weights to form a candidate set. Statistical analysis is performed on their sample mean μ and standard deviation σ. Based on the Raida criterion, data points with relative mean deviations exceeding twice the standard deviation are removed to ensure the centrality and credibility of the retained data. The remaining data is then filled into the knowledge graph using an arithmetic mean to avoid outliers misleading the capacity modeling results. This criterion, by limiting the deviation threshold range, effectively improves the stability and realism of entity attributes in the knowledge graph, providing high-quality support for downstream delivery time assessment and risk perception. Combined with the price sensitivity discussed in previous steps... With delivery urgency The assessment shows that this data aggregation mechanism further strengthens the data foundation for dynamic weight governance, enabling comprehensive weighting. The output results are traceable and consistent.

[0096] S400. The treated dataset is classified and stored according to industry scenarios. Specifically, the time-series database stores price fluctuation and inventory dynamic data; the graph database stores supply chain topology relationships; and the columnar database stores chip performance parameter sets.

[0097] In this embodiment, to address the management needs of time-series data such as chip price fluctuations and inventory dynamics, the system implements a storage optimization strategy of partition compression and multi-granularity aggregation in the time-series database. The time-series database is stored using this optimization strategy, specifically including:

[0098] The system partitions data based on the chip model field, and maintains a three-tiered data retention mechanism within each partition: High-frequency raw data is collected at 5-minute intervals, retaining the most recent 7 days; data is aggregated in 1-hour windows, calculating the mean, maximum, and minimum values, and retaining them for 30 days; price volatility is calculated daily.

[0099]

[0100] in, For volatility, The price of each transaction within the day. The system records this indicator as the daily average price, with a 1-day granularity and retains it for 2 years to support cross-cycle risk modeling and supply strategy evaluation.

[0101] A two-tiered indexing mechanism is constructed: the primary index uses a combination key of chip model and timestamp; the secondary index sets high-risk marking logic based on price change gradients, whereby any record meets the criteria... hour, The price at the previous moment is automatically labeled as "high-risk" for priority reading and processing by the dynamic weight management module. This strategy works in conjunction with the real-time calculation mechanism of price sensitivity Sp in step S300 to improve the visibility and processing priority of high-risk price signals in the system. It also provides data compression support and scheduling guarantee for delivery risk warning and alternative solution decision-making in step S500.

[0102] In this embodiment, the graph database serves as the core structure for storing supply chain topology relationships. Its relationship indexing mechanism achieves rapid parsing of upstream and downstream paths and accurate location of risk diffusion across the entire industry chain by constructing a bidirectional adjacency matrix. The graph database implements relationship indexing by constructing a bidirectional adjacency matrix, specifically through the following methods:

[0103] The adjacency matrix is ​​mapped along the path of "enterprise → product model → upstream raw material supplier" to build a vertical traceability path for the manufacturing chain; the column direction is mapped along the path of "enterprise → downstream customer → vehicle manufacturer" to track the distribution of chip products in end applications and order mapping relationships.

[0104] Introducing relational weight attributes, supply weight ,in, This indicates the percentage of the annual procurement amount for the corresponding raw materials or components. Indicates logistics dependence; substitution weight , For certification compatibility scoring, the score is based on whether the chip has passed the same certification body or holds a valid automotive-grade certificate. Mp represents the performance matching degree, which quantifies the similarity of the chips in terms of power consumption, temperature drift, and interface parameters.

[0105] Specifically, to enhance the quantitative expression of the connection strength between paths, the system introduces relationship weight attributes, including supply weight and substitution weight. Supply weight measures the impact of a particular raw material or component supply path on the company's normal operations, and is defined as follows:

[0106]

[0107] in, This indicates the proportion of the raw materials or components corresponding to this path in the annual procurement amount, with a value ranging from 0 to 1, reflecting the economic weight of this path in the overall procurement system; It represents the degree of logistics dependence. The value is based on the path stability index model. If there is a high dependence on a single logistics node or a high-risk area, the value will approach 1, which means that the path is highly vulnerable.

[0108] Alternation weights are used to evaluate the availability and substitutability of an alternative path, and are defined as follows:

[0109]

[0110] in, For certification compatibility scoring, the score is based on whether the chip has passed the same certification body or holds a valid automotive-grade certificate, with a maximum score of 1. To assess performance matching, the similarity of chips across parameters such as power consumption, temperature drift, and interfaces is quantified. This calculation involves edge weight modeling of chip substitution relationships in step S200. The closer the weight value is to 1, the more feasible the substitution path is, and the better it can mitigate the risk propagation during supply disruptions.

[0111] The bidirectional adjacency matrix and relation weight model together constitute the core mechanism of path indexing in the graph database. This not only supports the importance ranking of supply chain nodes in the dynamic weight governance of step S300, but also provides structured data support for path reconstruction and alternative scheduling in step S500. Based on these weights, the system can calculate the total impact strength of paths or construct a minimum-risk alternative subgraph, achieving efficient regulation and risk mitigation in complex supply chain networks.

[0112] In this embodiment, to address the efficiency issue of querying chip performance parameters, the system employs a scenario-oriented column family reconstruction and bitmap indexing mechanism in the columnar database to accelerate the process. First, based on the chip's application in different functional domains of new energy vehicles, the performance parameters are divided into three column families. The power domain column family focuses on the chip's high computing power and high reliability requirements in motor control and energy management, including key indicators such as maximum computing power, failure rate, and operating temperature range. The cockpit domain column family serves image display and human-machine interaction scenarios, including multimedia processing parameters such as GPU rendering frame rate and the number of multi-screen supports. The communication domain column family covers vehicle network-related performance, such as CANFD bus bandwidth and Ethernet communication latency, used to evaluate the chip's capabilities in high-real-time data exchange. This structured column family division is consistent with the multi-layered tagging system for chip models and automotive-grade applications in step S200, making the query path compact and the results focused.

[0113] In terms of the indexing mechanism, the system constructs a bitmap index for each column family, using the parameter value range as the segmentation standard, mapping different numerical ranges to binary marker bits. Taking the maximum computing power in the power domain as an example, when the maximum computing power of a chip is between [50, 100] DMIPS, the index marker bit is assigned 01; if the computing power value is in the range of [101, 200] DMIPS, the marker bit is 10. Temperature indicators are divided according to the difference between actual industrial-grade and automotive-grade standards: -40℃ to 125℃ corresponds to marker bit 00, and -40℃ to 150℃ corresponds to marker bit 11. This is achieved by constructing an index vector similar to the following:

[0114]

[0115]

[0116] The system can perform parallel matching of a large number of records through bitwise operations in complex filtering queries, significantly improving query efficiency in chip selection, performance comparison, and authentication verification. This indexing mechanism forms a data interface with the hierarchical evaluation of performance matching degree in the substitution weight in step S300, and also supports the rapid retrieval of chip models with equivalent or adjustable degradation solutions under high-pressure risk scenarios in step S500, providing structured and low-latency decision support for OEMs to supply alternatives.

[0117] S500. A supply chain risk warning is triggered based on stored data, outputting the probability of delivery delay and alternative chip solutions. The triggering mechanism for the supply chain risk warning is based on a multi-factor fusion probability model and tiered alternative solution recommendations. In this embodiment, the specific method for triggering a supply chain risk warning based on stored data and outputting the probability of delivery delay and alternative chip solutions includes:

[0118] Calculate the comprehensive risk factor: ;in, This is an inventory warning flag, with a value of 0 or 1, indicating whether the inventory is below the safety threshold. This is a capacity overload indicator, with a value ranging from 0 to 1, representing the degree of production line load. This is a logistics risk value, reflecting the stability of the route and the level of external risk.

[0119] Introducing the Sigmoid function to calculate the probability of delivery delay: , where K is a positive real parameter that adjusts the steepness of the curve;

[0120] A three-tiered chip matching system is implemented: Level 1 matching screens chip models with power consumption and operating voltage deviations within ±5%; Level 2 matching prioritizes models with fully pin-to-pin compatibility, and if none are available, adapter boards are permitted with adapter costs limited to $0.20; Level 3 matching prioritizes chips with the same certification certificate number. This three-tiered matching system provides enterprises with multi-level, industry-standard alternative chip solutions, effectively mitigating supply chain risks and ensuring stable chip supply.

[0121] Specifically, after an alert is triggered, the system performs a three-tiered alternative chip matching process to mitigate supply chain risks. The first tier focuses on core electrical performance parameters, screening chip models with power consumption and operating voltage deviations within ±5%, ensuring the approximation of alternatives in terms of energy consumption and circuit compatibility, and enhancing the reliability of the replacement. The second tier prioritizes models with fully pin-to-pin compatible packaging, prioritizing those without a complete match. If no fully compatible model is available, an adapter board solution is permitted, with the adapter cost limited to $0.20, ensuring the economic rationality of physical assembly adjustments. The third tier considers automotive-grade certification status, prioritizing chips with the same certification certificate number to ensure the compliance and market access requirements of the alternative solutions. This three-tiered matching system relies on the knowledge graph built in step S200 and the performance parameter database in step S400 to achieve efficient and industry-standard alternative solution selection. Combined with real-time adjustment of dynamic weights in step S300, it ensures the accuracy and feasibility of the alert response, ultimately supporting the stable operation and adaptive risk management of the new energy vehicle chip supply chain.

[0122] The multi-dimensional heterogeneous data acquisition and dynamic governance system for the new energy vehicle chip industry of the present invention has the following beneficial effects:

[0123] 1. Comprehensive data collection to ensure data timeliness and comprehensiveness: Through the distributed data pipeline in step S100, data is collected in real time from multiple channels such as EDA toolchain, wafer fab MES system, supply chain SaaS platform, and market transaction database, covering the entire chain of chip design, production, supply chain, and market. Compared with the traditional single data source collection method, the data coverage is more comprehensive and the collection efficiency is higher, ensuring the real-time nature of the data and providing a high-quality data foundation for subsequent data processing and application.

[0124] 2. Dynamic weight governance enhances data credibility and accuracy: The dynamic weight model constructed in step S300 combines price sensitivity and delivery urgency, and adjusts adaptive coefficients in real time based on the market supply and demand ratio to achieve precise cleaning of conflicting data. At the same time, through a data aggregation mechanism, high-weight data is retained and low-weight data is removed. Compared with the traditional fixed-weight governance method, data governance is more in line with market dynamics, conflicting data is processed more accurately, and data credibility is significantly improved, providing reliable data support for industry decision-making.

[0125] 3. Knowledge Graph Support for Clearly Defining Industry Relationships and Risk Paths: The chip industry knowledge graph constructed in step S200 adopts a four-layer structure to organize entity relationships, marking chip substitution relationships and supply chain risk transmission paths. This solves the problems of chaotic industry relationships and difficulty in risk tracking under traditional industry relationship analysis methods. It can clearly present the entire chain of industry relationships from "design to manufacturing to packaging and testing to certification", accurately track the supply chain risk transmission paths, and provide structured knowledge support for risk warning and alternative solution generation.

[0126] 4. Categorized storage optimization to improve data storage efficiency and query performance: Step S400 categorizes and stores the treated data into time-series databases, graph databases, and columnar databases according to scenarios. Optimization mechanisms such as partition compression, bidirectional adjacency matrix, and bitmap indexes are adopted for each database. Compared with the traditional single database storage method, data storage is more reasonable, query efficiency is greatly improved, and it can quickly respond to various data application needs such as price calculation, risk analysis, and chip selection.

[0127] 5. Precise risk warning and alternative solutions to ensure supply chain stability: Step S500 calculates the probability of delivery delay based on a multi-factor fusion probability model to achieve precise early warning of supply chain risks; at the same time, it performs three-level alternative chip matching based on knowledge graphs and databases to quickly provide alternative solutions that meet the requirements of electrical performance, packaging form and automotive-grade certification, which solves the problems of delayed warning and lack of alternative solutions under traditional risk warning methods, effectively alleviates supply chain risks and ensures the stable operation of the new energy vehicle chip supply chain.

[0128] Based on the same inventive concept, embodiments of the present invention also provide an electronic device. Figure 2 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Figure 2 As shown, an embodiment of the present invention provides an electronic device including: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, which, when executed by the one or more processors, enable the one or more processors to implement any of the acquisition and management methods described in the above embodiments; the one or more I / O interfaces 103 are connected between the processors and the memory, configured to enable information interaction between the processors and the memory.

[0129] The processor 101 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 102 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and flash memory (FLASH); the I / O interface (read / write interface) 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus (Bus).

[0130] In some embodiments, the processor 101, memory 102, and I / O interface 103 are interconnected via bus 104, and thus connected to other components of the computing device.

[0131] In some embodiments, the one or more processors 101 include a field-programmable gate array.

[0132] This invention also provides a computer-readable medium. The computer-readable medium stores a computer program, which, when executed by a processor, implements the steps of any of the acquisition and management methods described in the above embodiments. The computer-readable storage medium can be volatile or non-volatile.

[0133] This invention also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in the processor of an electronic device, the processor in the electronic device executes the above-described acquisition and management method.

[0134] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0135] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0136] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0137] The computer program instructions used to perform the operations of this invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions. This electronic circuitry can execute the computer-readable program instructions to implement various aspects of the invention.

[0138] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0139] Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0140] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0141] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0142] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0143] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of the invention as set forth in the appended claims.

Claims

1. A method for collecting and managing data for new energy vehicle chip industry, characterized in that, Comprise: Real-time collection of new energy vehicle chip industry data through a distributed data pipeline; specifically including: obtaining design parameters including power consumption, computing power, and temperature drift coefficient from EDA tool chain; extracting capacity utilization, yield, and work-in-process status from wafer factory MES system; synchronizing order delivery time, inventory depth, and logistics node information from supply chain SaaS platform; capturing price fluctuation time series data and enterprise public opinion events from market transaction database; Based on the chip industry data, a chip industry knowledge graph is constructed, and enterprise entities, chip product models, and supply chain nodes are associated as a topological network, wherein the enterprise entities include design companies, wafer factories, packaging and testing factories, and automotive certification agencies; the chip product models are associated with multi-level BOM lists and substitution relationship matrices; the supply chain nodes are labeled with geographic coordinates, transportation routes, and capacity elasticity coefficients; Dynamic weight governance is performed on the knowledge graph to obtain a weight-governed dataset; specifically including: loading a dynamic weight model for each data record in the knowledge graph, the model calculation rule is: weight value W = α × price sensitivity + β × delivery urgency, where α and β are adaptive coefficients adjusted in real time according to market supply and demand ratio; based on the weight value W, conflicting data is cleaned, data with weight value W greater than threshold T is retained, and cross-source data is aggregated to generate a governed dataset; The governed dataset is stored according to industry scenarios, specifically, time series database stores price fluctuation and inventory dynamic data; graph database stores supply chain topological relationship; columnar database stores chip performance parameter set; Based on the stored data, supply chain risk early warning is triggered, and delivery delay probability and substitute chip scheme are output.

2. The method of claim 1, wherein, The entity relationship chain of the chip industry knowledge graph is constructed using a four-layer structure: The first layer takes the design company ID as the root node, connects the chip model set bound to it, and forms a master-slave mapping relationship between design source and product model; The second layer maps each chip model to the corresponding wafer manufacturing process node and the required mask ID, forming a vertical expansion of manufacturing parameters; The third layer takes the wafer process node as an intermediate entity, associates the packaging and testing factory number and key solder material code, and realizes information extension from the process to the packaging and testing link; The fourth layer takes the packaging and testing factory as the starting point, connects the automotive certification certificate number and its expiration date, and supports traceability management of compliance and certification timeliness.

3. The method of claim 1, wherein, The establishment conditions of chip model substitution relationship include: The power consumption deviation of two chip models satisfies PA−PB / PA≤0.1, PA and PB are the rated power consumptions of chip A and chip B respectively, and the temperature drift coefficients are consistent; two chip models have consistency in automotive certification path, satisfying sharing the same certification certificate number or having the same certification agency; after satisfying the above conditions, an alternative edge is established between the two chip models in the knowledge graph, the edge has directionality and conditional constraint attribute, and the edge weight participates in the calculation of the dynamic weight model.

4. The method of claim 1, wherein, Based on the upstream and downstream traversal mechanism of multi-level entity relationship in the knowledge graph, the dynamic labeling of supply chain risk transmission path is realized, and the specific method includes: When a production interruption, equipment failure, or geopolitical event is detected in a wafer factory, a three-level recursive analysis of relevant nodes in the knowledge graph is triggered: the first level analyzes key chip models whose production capacity supplied by the wafer factory accounts for more than a preset proportion in a certain chip model; the second level analyzes the position of the affected chip model in the BOM list of the vehicle factory and the corresponding order quantity; the third level analyzes chip models that meet the replaceable conditions and their current inventory locations in the test and packaging factory; Assigning an impact strength value to each risk transmission path , the calculation formula is , wherein E is the event level, which is classified according to the wafer factory downtime: E=1 when the downtime is less than 3 days, E=2 when the downtime is between 3 and 7 days, and E=3 when the downtime exceeds 7 days, reflecting the basic severity of the event; Rs is the inventory satisfaction rate of the replacement chip, and the calculation method is: ; the path with a higher response strength than the preset threshold is dynamically marked in the knowledge graph with a preset color.

5. The method of harvesting and governing of claim 1, wherein, The calculation process of the price sensitivity includes: Taking the preset time as the monitoring window, the time series data of the chip price is intercepted from the market transaction database, the price curve is fitted by using the cubic spline interpolation method, and the price change amount per unit time is calculated wherein, ΔP represents the price change amount between the start and end points of the window of the fitted price curve, with unit of yuan; Δt represents the time span of the monitoring window, fixed as 5 minutes; Extract volume data in the same monitoring window, calculate the standard deviation of volume σV, divide by the average volume , get the volume volatility , A public opinion event compensation factor γ is set, when it is detected that the preset risk words are contained in the text in the monitoring window, γ = 1.5, otherwise γ = 1; the calculation formula of the price sensitivity is: 。 6. The method of harvesting and governing of claim 1, wherein, The calculation process of the delivery urgency includes: extracting a standard lead time corresponding to a target chip model from a knowledge graph and a current or predicted capacity utilization of a wafer fab associated with the chip ; Dynamic inventory depth is calculated based on inventory information obtained from the supply chain platform and recent consumption data , where , is the current inventory level is the average daily shipment or consumption over the past 7 days; if days, activate the inventory alert flag urgency of the due date The calculation is made using the formula: ； where the denominator represents the combined effect of inventory and capacity supply capabilities, which are normalized to the interval [0.1, 10].

7. The method of harvesting and governing of claim 1, wherein, The aggregated cross-source data generates a governed data set, and the specific method includes: Aggregate inventory information for the same chip model, get a set of raw inventory values from k independent data sources , filter according to a set of dynamic weights , keep data sources, effective inventory calculation formula: ； The wafer factory production capacity data is subjected to consistency verification, the three data sources with the highest weight values are selected to form a candidate set, the sample mean μ and standard deviation σ are calculated, data points with a relative mean deviation exceeding twice the standard deviation are removed based on the Laplace criterion, and the remaining data is filled with arithmetic mean to the knowledge graph.

8. The method of harvesting and governing of claim 1, wherein, The time series database adopts a partition compression and multi-granularity aggregation optimization strategy for storage, specifically including: According to the chip model field, each partition maintains a three-level data retention mechanism: high-frequency raw data is collected with a granularity of 5 minutes, and the last 7 days are retained; data is aggregated with a window of 1 hour, the mean, maximum, and minimum values are calculated, and 30 days are retained; the price volatility is calculated by day: ； wherein, is the volatility, is the price of each trade within the day, is the daily average price, the system records this metric at 1-day granularity and keeps it for 2 years to support cross-period risk modeling and supply strategy evaluation; A double-layer index mechanism is constructed: the first-level index adopts the combination key of chip model and timestamp; the second-level index sets high-risk marker logic based on price change gradient. When any record meets , the price of the previous moment is automatically labeled as "high risk".

9. The method of harvesting and governing of claim 1, wherein, The graph database realizes relationship indexing by constructing a bidirectional adjacency matrix, and the specific method includes: The adjacency matrix row direction is mapped along the "enterprise→product model→upstream raw material supplier" path to construct a vertical traceability path of the manufacturing chain; the column direction is mapped along the "enterprise→downstream customer→vehicle factory" direction to track the distribution and order mapping relationship of chip products in terminal applications; Introducing relationship weight attribute, supply weight wherein, represents the proportion of the path corresponding to the raw material or component in the annual procurement amount, represents the logistics dependence degree; substitution weight , is an authentication compatibility score, which is scored according to whether the chip passes the same authentication agency or holds a valid vehicle regulation certificate, and Mp is a performance matching degree, which quantifies the similarity of the chip in the power consumption, temperature drift and interface parameter dimensions.

10. The method of harvesting and governing of claim 1, wherein, Based on the stored data, the supply chain risk early warning is triggered, and the delivery delay probability and alternative chip scheme are output, and the specific method includes: The calculation of the comprehensive risk factor is as follows: ; wherein, is a stock early warning mark, and has a value of 0 or 1, representing whether the stock is lower than a safety threshold; is a capacity overload mark, and has a value ranging from 0 to 1, representing the degree of load of the production line; is a logistics risk value, reflecting path stability and external risk level; The sigmoid function is introduced to calculate the probability of delay at the intersection: where K is a positive real number parameter that adjusts the steepness of the curve. Three-level alternative chip matching is performed: first-level matching selects chip models with a power consumption and working voltage deviation within a preset range; second-level matching preferentially selects models with completely compatible pins, and if there are none, a conversion board scheme is allowed with a conversion cost limited within a preset price; third-level matching preferentially recommends chips with the same certification certificate number.

Images