Engineering material supply chain risk early warning method and system based on multi-source data fusion

Abstract

The present application relates to the technical field of supply chain risk early warning, and discloses an engineering material supply chain risk early warning method and system based on multi-source data fusion, which comprises the following steps: performing multi-dimensional semantic analysis on engineering supply correlation information to obtain a standard record set; performing loss weighting on the standard record set based on material aging characteristics and maintenance deviation to obtain a decay record set; performing attribute analysis on the decay record set to obtain feature descriptors of the same supply batch; performing vulnerability identification on the decay record set based on the mapping intersection relationship between the origin substitution constraint and the mapping relationship between the current bid section pile number interval and the mix ratio number to obtain a dependent marker group; taking the irreplaceable origin of the marker in the dependent marker group as an anchor point, and performing backtracking and optimization on the historical supply record to obtain a substitution topology graph; and performing strategy matching on the substitution topology graph to obtain a supply guarantee instruction set. The present application can improve the efficiency of engineering material supply chain risk early warning based on multi-source data fusion.

Description

Technical Field

[0001] This invention relates to the field of supply chain risk early warning technology, and in particular to a method and system for early warning of supply chain risks in engineering materials based on multi-source data fusion. Background Technology

[0002] Existing technologies largely rely on isolated business systems for information collection, resulting in severe semantic heterogeneity between different data interfaces. This makes it difficult to integrate cross-system engineering supply-related information into a structurally consistent analytical object. Furthermore, existing methods, when assessing material quality status, typically rely solely on factory records or single sampling results, lacking a dynamic quantification mechanism for performance degradation caused by time effects and environmental fluctuations during transportation and storage. This leads to significant deviations in the timeliness and accuracy of the data foundation used for risk identification, failing to accurately reflect the immediate usability of batch materials.

[0003] Current technologies for identifying supply chain risks often only conduct general assessments of supplier qualifications or historical fulfillment rates, failing to deeply correlate the material's own origin substitution constraints with the technical requirements of specific sections and mix proportions at the engineering site. This coarse-grained assessment model neglects the precise identification of vulnerable links with a unique dependence on a single origin. Consequently, when disturbances occur at the source, the generation of alternative supply paths lacks specificity and operability. The overall early warning strategy can only provide macro-level situational awareness, making it difficult to issue refined assurance instructions targeting specific, irreplaceable origins. Ultimately, this results in delayed risk response and insufficient engineering continuity assurance capabilities. Therefore, improving the efficiency of risk early warning in the engineering materials supply chain has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method and system for risk early warning of engineering materials supply chain based on multi-source data fusion, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for risk early warning of engineering materials supply chain based on multi-source data fusion, comprising: P1. Perform multi-dimensional semantic parsing on engineering supply association information originating from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information; P2. Based on the material aging characteristics and maintenance deviations implicit in the standard record set, the standard record set corresponding to the same supply batch is weighted for depreciation to obtain the depreciation record set of the same supply batch. P3. Perform attribute parsing on the attenuation record set to obtain the feature descriptor of the same supply batch; P4. Based on the origin substitution constraint of the feature descriptor and the mapping intersection relationship between the current section chainage interval and the mix ratio number, the vulnerability of the attenuation record set is identified, and the dependency label group of the attenuation record set is obtained. P5. Using the non-substitutable production sites marked in the dependency tag group as anchor points, backtrack and select the best historical supply records to obtain a substitution topology map oriented towards non-substitutable production sites. P6. Perform strategy matching on the alternative topology map to obtain a supply guarantee instruction set for irreplaceable production areas.

[0006] In a preferred embodiment, multi-dimensional semantic parsing is performed on engineering supply association information originating from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information, including: Field decomposition is performed on engineering supply association information originating from multiple heterogeneous data interfaces to obtain the original field set of engineering supply association information; Based on semantic mapping rules, the value range of the original field set is transformed to obtain the normalized field set of the original field set; The normalized field set is reconstructed to obtain a standardized record set of engineering supply-related information.

[0007] In a preferred embodiment, based on the material aging characteristics and maintenance deviations implicit in the specification record set, a depreciation weighting is applied to the specification record set corresponding to the same supply batch to obtain a depreciation record set for the same supply batch, including: Time effect analysis is performed on the standard record set to obtain the set of attenuation parameters for the same supply batch; Obtain monitoring temperature and humidity records for the same supply batch during transportation and storage to obtain environmental measurement data for the same supply batch; Based on the attenuation parameter set and environmental measurement data, the descriptive items reflecting the material quality status in the standard record set are weighted by loss to obtain the attenuation record set of the same supply batch.

[0008] In a preferred embodiment, based on the attenuation parameter set and environmental measurement data, the confidence level of the descriptive items reflecting the material quality status in the specification record set is adjusted to obtain an attenuation record set for the same supply batch, including: Parametric extraction is performed on the attenuation parameter set to obtain the attenuation rate of the same supply batch; Accelerated extrapolation of environmental measurement data yields the environmental acceleration factor for the same supply batch. The duration of each descriptive item that reflects the material quality status in the standard records is obtained. Initial confidence levels are extracted for each descriptive term to obtain the initial confidence level for each descriptive term. Then, confidence discounting is performed on each descriptive term according to the following formula to obtain the adjusted confidence level for each descriptive term: ; In the formula, As the reference decay rate, As an environmental accelerator, The duration is... As the initial confidence level, To adjust the confidence level; The adjusted confidence levels are aggregated to obtain a set of attenuation records for the same supply batch.

[0009] In a preferred embodiment, attribute parsing is performed on the attenuation record set to obtain feature descriptors for the same supply batch, including: Extract core attribute fields from the attenuation record set and form a core field group for the same supply batch; The values ​​of related fields within the core field group are merged and aggregated to form aggregated attribute items for the same supply batch; The aggregated attribute items are described and encapsulated to obtain the feature descriptors for the same supply batch.

[0010] In a preferred embodiment, based on the origin substitution constraint of the feature descriptor and the mapping intersection relationship between the current section chainage interval and the mix proportion number, vulnerability identification is performed on the attenuation record set to obtain a dependency tag group for the attenuation record set, including: Constraint fields are extracted from the feature descriptors to obtain the origin identifier and blending ratio number for the same supply batch; Based on the mapping and intersection relationship between the current section chainage interval and the mix proportion number, determine whether there is a compliant alternative origin other than the origin identifier for the mix proportion number; When it is determined that there is no compliant alternative origin, the origin identifier is marked as an irreplaceable origin and merged into the dependency tag group of the decay record set.

[0011] In a preferred embodiment, based on the mapping intersection between the current lot number interval and the mix proportion number, it is determined whether the mix proportion number has a compliant alternative origin other than the origin identifier, including: Query the mix proportion for the current chainage interval to obtain the mix proportion number corresponding to the current chainage interval; By binding the corresponding blending ratio number to the place of origin for traceability, a set of compliant places of origin for the corresponding blending ratio number is obtained; By comparing the differences in the compliant origin set, if the compliant origin set only contains the origin identifier, it is confirmed that there is no compliant alternative origin.

[0012] In a preferred embodiment, using the non-substitutable production sites marked in the dependency tagging group as anchor points, historical supply records are backtracked and optimized to obtain a substitution topology map oriented towards non-substitutable production sites, including: By performing correlation filtering on historical supply records, a subset of related records with supply relationships to irreplaceable production areas is obtained; Perform blending ratio matching on a subset of associated records to obtain alternative supply source records that meet the blending ratio number corresponding to the irreplaceable production area; The records of the candidate supply sources are sorted by impedance to obtain a hierarchy of merits for each candidate supply source. Based on the hierarchy of superiority and inferiority, a topology is constructed for the reachable paths between alternative supply source records and irreplaceable production sites to obtain an alternative topology map oriented towards irreplaceable production sites.

[0013] In a preferred embodiment, strategy matching is performed on the alternative topology map to obtain a supply security instruction set for non-substitutable production locations, including: Structural analysis is performed on the alternative topology graph to obtain its topological characteristic parameters; Threshold comparison of topological feature parameters yields risk level identifiers for alternative topological maps; By mapping risk level identifiers to instructions, a set of supply guarantee instructions for irreplaceable production areas is obtained.

[0014] To address the aforementioned problems, this invention also provides an engineering materials supply chain risk early warning system based on multi-source data fusion, used to execute the engineering materials supply chain risk early warning method based on multi-source data fusion described in the above embodiments. The system includes: The semantic parsing module is used to perform multi-dimensional semantic parsing on engineering supply association information from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information; The loss weighting module is used to assign loss weights to the standard record set corresponding to the same supply batch based on the material aging characteristics and maintenance deviations implicit in the standard record set, so as to obtain the attenuation record set of the same supply batch. The attribute parsing module is used to parse the attributes of the attenuation record set to obtain the feature descriptors of the same supply batch. The vulnerability identification module is used to identify the vulnerability of the attenuation record set based on the origin substitution constraint of the feature descriptor and the mapping intersection relationship between the current section chainage interval and the mix ratio number, and obtain the dependency tag group of the attenuation record set. The backtracking optimization module is used to backtrack and optimize historical supply records using the non-substitutable production areas marked in the dependency tag group as anchor points, so as to obtain an alternative topology map oriented towards non-substitutable production areas. The strategy matching module is used to perform strategy matching on the alternative topology map to obtain a set of supply guarantee instructions for irreplaceable production locations.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention reconstructs the engineering supply association information of heterogeneous data interfaces through multidimensional semantic parsing, and introduces material aging characteristics and maintenance deviations to assign depreciation weights to the record set, thereby improving the dynamic accuracy of material quality status assessment. The method quantifies the time effects and environmental fluctuations during transportation and storage as attenuation factors for record confidence, ensuring that the data used for risk warning accurately reflects the immediate usability of batch materials. This enhances the ability to detect material performance degradation and effectively improves the timeliness and accuracy of risk detection.

[0016] 2. This invention identifies vulnerabilities based on the origin identifier in the feature descriptor, combined with the mapping intersection between the current section chainage interval and the mix proportion number. It accurately locates dependent marker groups with no compliant alternative origins and uses these as anchor points to trace back historical supply records to generate an alternative topology map. By identifying the risk level and mapping instructions to the topology structure, it outputs a supply guarantee instruction set for irreplaceable origins, enhancing the planning efficiency and operability of alternative supply routes and significantly improving the rapid response and continuous guarantee capability of the engineering material supply chain when key nodes are constrained. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a method for early warning of engineering material supply chain risks based on multi-source data fusion, provided in an embodiment of the present invention; Figure 2 A functional module diagram of an engineering materials supply chain risk early warning system based on multi-source data fusion provided in an embodiment of the present invention; Figure 3 A bar chart showing the trigger frequency of the supply assurance instruction set provided in an embodiment of the present invention; Figure 4 This is a supply chain topology risk demarcation curve provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a method for risk warning of engineering materials supply chain based on multi-source data fusion. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for risk warning of engineering materials supply chain based on multi-source data fusion can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating a method for early warning of engineering material supply chain risks based on multi-source data fusion, according to an embodiment of the present invention. In this embodiment, the method for early warning of engineering material supply chain risks based on multi-source data fusion includes: P1. Perform multi-dimensional semantic parsing on engineering supply association information originating from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information.

[0021] In this embodiment of the invention, the step of performing multi-dimensional semantic parsing on engineering supply association information originating from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information includes: Field decomposition is performed on engineering supply association information originating from multiple heterogeneous data interfaces to obtain the original field set of engineering supply association information; Based on semantic mapping rules, the value range of the original field set is transformed to obtain the normalized field set of the original field set; The normalized field set is reconstructed to obtain a standardized record set of engineering supply-related information.

[0022] When decomposing the fields of engineering supply association information from multiple heterogeneous data interfaces, the original data body of the engineering supply association information is read from each interface. According to the data structure definition agreed upon by each interface, all field identifiers and their corresponding field contents in the data body are identified and extracted. All field identifiers and field contents from different interfaces are aggregated into a unified set as independent items to obtain the original field set of engineering supply association information.

[0023] Based on a pre-built semantic mapping rule set, the original field set is transformed in terms of value range. This semantic mapping rule set consists of multiple mapping rules. Each rule explicitly specifies the correspondence between an original field name and a standard field name, as well as a conversion lookup table between the original field value and the standard field value. During processing, each item in the original field set is traversed to check if there is a rule in the mapping rules that matches the original field name. If a matching rule exists, the field name of the item is replaced with the standard field name in the rule, and the field content is converted into the corresponding standard field value according to the conversion lookup table. If no matching rule exists, the original field name and field content of the item are retained unchanged. After processing, all items are merged to form a normalized field set.

[0024] When reconstructing records in a normalized field set, the first step is to use the standardized record structure template defined by the business object. This template specifies the complete set of fields that make up a standardized record. In the normalized field set, the business primary key identifier is used as the grouping basis. All field items belonging to the same business record are grouped into one group. Each group corresponds to one record to be assembled. For each group, the fields are aligned one by one according to the field set in the template. If the group already contains the field specified in the template, the field value is directly filled in. If the group is missing a field specified in the template, the field is added and a default value is assigned. If there are redundant fields in the group that are not specified in the template, they are discarded. After assembly, a set of standardized records with a unified structure and complete fields is obtained. All records together constitute a standardized record set of engineering supply-related information.

[0025] The beneficial effects are as follows: By reading the data body of engineering supply-related information from multiple heterogeneous data interfaces one by one and decomposing the fields according to the data structure definition of each interface, the field identifiers and field contents from different sources are uniformly aggregated into an original field set, eliminating information dispersion and format fragmentation caused by interface differences. Based on a pre-built semantic mapping rule set, the original field set is transformed in terms of value range. The correspondence between original field names and standard field names, as well as the field value conversion lookup table, are used to unify the names and standardize the content of all items, ensuring semantic consistency of subsequent processed objects. During the record reconstruction process, the normalized field set is grouped and aligned based on the complete field set specified by the standardized record structure template. Field items belonging to the same business record are merged based on the business primary key. Missing fields are filled in according to the template and assigned default values ​​while redundant fields are discarded, resulting in a standardized record set with a unified structure and complete fields. This provides a high-quality structured data foundation that can be directly calculated for material aging characteristic analysis and loss weighting.

[0026] P2. Based on the material aging characteristics and maintenance deviations implicit in the standard record set, the standard record set corresponding to the same supply batch is weighted for depreciation to obtain the depreciation record set of the same supply batch.

[0027] In this embodiment of the invention, the step of assigning depreciation weights to the standard record set corresponding to the same supply batch based on the material aging characteristics and maintenance deviations implicit in the standard record set to obtain the depreciation record set of the same supply batch includes: Time effect analysis is performed on the standard record set to obtain the set of attenuation parameters for the same supply batch; Obtain monitoring temperature and humidity records for the same supply batch during transportation and storage to obtain environmental measurement data for the same supply batch; Based on the attenuation parameter set and environmental measurement data, the descriptive items reflecting the material quality status in the standard record set are weighted by loss to obtain the attenuation record set of the same supply batch.

[0028] Based on the attenuation parameter set and environmental measurement data, the confidence level of the descriptive items reflecting the material quality status in the specification record set is adjusted to obtain the attenuation record set for the same supply batch, including: Parametric extraction is performed on the attenuation parameter set to obtain the attenuation rate of the same supply batch; Accelerated extrapolation of environmental measurement data yields the environmental acceleration factor for the same supply batch. The duration of each descriptive item that reflects the material quality status in the standard records is obtained. Initial confidence levels are extracted for each descriptive term to obtain the initial confidence level for each descriptive term. Then, confidence discounting is performed on each descriptive term according to the following formula to obtain the adjusted confidence level for each descriptive term: ; In the formula, As the reference decay rate, As an environmental accelerator, The duration is... As the initial confidence level, To adjust the confidence level; The adjusted confidence levels are aggregated to obtain a set of attenuation records for the same supply batch.

[0029] When performing time effect analysis on the standard record set, all records bound to the same supply batch are extracted one by one from the standard record set. The record generation timestamp and the material manufacturing time recorded in each record are read. The time span between the record generation timestamp and the material manufacturing time is calculated as the duration of the batch. At the same time, the theoretical performance retention years under standard curing conditions of the material are extracted from the records. The duration of the batch is matched with the theoretical performance retention years item by item to generate the proportional relationship of the decay of various material properties with time under standard conditions and the remaining effective duration. These proportional relationships and the remaining effective duration are uniformly compiled into a set of decay parameters for the same supply batch.

[0030] When acquiring monitoring temperature and humidity records for the same supply batch during transportation and storage, temperature and humidity values ​​continuously collected over time are read from temperature and humidity sensing nodes deployed on transport vehicles and storage facilities. These temperature and humidity sequences arranged chronologically from the monitoring points are associated with transportation stage identifiers and storage stage identifiers. The validity of the temperature and humidity values ​​at each time point is verified to eliminate abnormal readings that exceed the sensing range. The valid temperature and humidity data that pass the verification are retained and arranged in the order of collection time to form continuous environmental time-series data. This environmental time-series data is used as the environmental measured data for the same supply batch.

[0031] When extracting parameters from the attenuation parameter set, the attenuation ratio and corresponding remaining effective duration of each material characteristic in the batch are extracted one by one from the attenuation parameter set. The attenuated amount of each material characteristic is compared with the duration of existence, and the performance attenuation amount of the material characteristic under standard conditions per unit time is calculated to obtain the attenuation rate of each material characteristic. The attenuation rates of all material characteristics together constitute the attenuation rate of the same supply batch.

[0032] When performing accelerated extrapolation on environmental measurement data, the environmental measurement data of the same supply batch are scanned sequentially according to the collection time. The duration and magnitude of temperature values ​​exceeding the standard maintenance temperature range and humidity values ​​exceeding the standard maintenance humidity range are statistically analyzed segment by segment. The magnitude of temperature and humidity exceeding the limits are weighted and synthesized with preset temperature and humidity acceleration influence coefficients to obtain the environmental stress intensity of the transportation and storage stage. Then, the environmental stress intensity and the duration of exceeding the limits are comprehensively extrapolated to obtain the acceleration factor of material performance degradation in this stage, which is used as the environmental acceleration factor for the same supply batch.

[0033] When obtaining the duration of each descriptive item reflecting the material quality status in the standard record set, the descriptive items marked as reflecting the material quality status in the standard record set are located one by one. The inspection timestamp or data generation timestamp associated with the descriptive item is read from the record corresponding to each descriptive item. The time length elapsed from the inspection or generation time to the current early warning analysis time of the descriptive item is calculated. The obtained time length is the duration of the descriptive item.

[0034] When extracting the initial confidence level for each descriptive item, the data confidence level label value attached to the descriptive item when it was generated is read from each descriptive item that reflects the material quality status. This confidence level label value can be the reliability level of the inspection report, the accuracy level of the testing instrument, or the credibility score of the data acquisition channel. If the confidence level is not labeled in the descriptive item, the default confidence level assignment for this type of descriptive item is searched from the metadata segment of the specification record set, and the confidence level value read or found is used as the initial confidence level for the descriptive item.

[0035] When performing confidence reduction calculations on a descriptive item, the obtained decay rate of the batch is matched with the decay rate of the material properties to which the corresponding descriptive item belongs to obtain the reference decay rate applicable to the descriptive item. The reference decay rate is then continuously multiplied by the environmental acceleration factor and the duration of the descriptive item to obtain the comprehensive decay index. Using the natural constant as the base and the negative value of the comprehensive decay index as the exponent, the decay discount coefficient is obtained by exponential calculation. The initial confidence level of the descriptive item is multiplied by the decay discount coefficient to reduce the loss of confidence of the descriptive item caused by the passage of time and environmental deviations. The product result is the adjusted confidence level of the descriptive item.

[0036] When aggregating records based on adjusted confidence levels, a correspondence is established between each descriptive item reflecting the material quality status and its calculated adjusted confidence level. The original field identifiers and field values ​​of the descriptive items are retained, and the adjusted confidence level is attached to the descriptive item as a weight label. All descriptive items adjusted by confidence level are grouped according to the identifier of the same supply batch, forming a set of records with depreciation weights for all quality status descriptive items in that batch. This set is the depreciation record set for the same supply batch.

[0037] The baseline decay rate is derived from the decay rate of the same supply batch obtained when extracting parameters from the decay parameter set. The decay rate is calculated by extracting the decay ratio and remaining effective duration of various material properties from the decay parameter set, comparing the decayed amount with the remaining duration, and calculating the performance decay per unit time under standard conditions. In application, the decay rate corresponding to the material property of the descriptive item is used as the baseline decay rate. The environmental acceleration factor is derived from the environmental acceleration extrapolation of environmental measurement data for the same supply batch. The environmental acceleration factor is calculated by statistically analyzing the temperature and humidity exceedance ranges and durations, and combining this with a preset acceleration influence coefficient to estimate the acceleration factor of material performance decay during transportation and storage. The duration is derived from the duration of each descriptive item reflecting the material quality status in the specification record set. The duration is obtained by reading the inspection timestamp associated with the descriptive item and calculating the time elapsed until the early warning analysis time. The initial confidence level comes from the initial confidence level of each descriptive item obtained when extracting the initial confidence level of each descriptive item. The initial confidence level is the data confidence level label value read from the descriptive item or the default confidence level assignment value found from the data segment of the canonical record set.

[0038] The confidence reduction calculation calculates the adjusted confidence level by multiplying the initial confidence level by an exponential decay term with the natural constant as the base. The exponent in the exponential decay term is the negative value of the product of the baseline decay rate, the environmental acceleration factor, and the duration. The product result comprehensively reflects the combined effect of the material's inherent decay characteristics, the intensity of environmental stress, and the time span. The essence of the confidence reduction calculation is to continuously discount the original credibility of the descriptive item. The discount ratio is jointly determined by the natural decay rate of the material property, the acceleration effect of the actual storage and transportation environment, and the aging degree of the data. The final output adjusted confidence level represents the degree to which the descriptive item is still acceptable after eliminating the effects of time effect and maintenance deviation.

[0039] When the duration is zero, the exponential decay term is one, and the adjusted confidence level equals the initial confidence level without any loss. As the duration increases, the exponential decay term gradually decreases and approaches zero, and the adjusted confidence level decreases monotonically and approaches zero. The higher the baseline decay rate, the faster the adjusted confidence level decays with the duration. The larger the environmental acceleration factor, the faster the adjusted confidence level decays with the duration. The adjusted confidence level always remains non-negative and does not exceed the trajectory of the initial confidence level.

[0040] The beneficial effects are as follows: By analyzing the time effect of the standard record set, each record bound to the same supply batch is extracted and its duration is calculated. Simultaneously, the attenuation ratio relationship is generated by combining the theoretical performance retention period with the remaining effective duration to form a set of attenuation parameters. This transforms the natural degradation trend of material performance over time into quantifiable attenuation parameters. Monitoring records are obtained from temperature and humidity sensors in transportation vehicles and storage facilities, and abnormal readings are removed to form environmental measured data. Parametric extraction is performed on the attenuation parameter set to obtain the attenuation rate of each material characteristic. Accelerated extrapolation is then performed on the environmental measured data to obtain the environmental acceleration factor, incorporating the additional accelerated damage to material performance caused by deviations in storage and transportation environments into the quantitative framework. Initial confidence levels and durations are extracted for each descriptive item reflecting the material quality status. The baseline decay rate, environmental acceleration factor, and duration are continuously multiplied and exponentially calculated to obtain the decay discount coefficient. The adjusted confidence level is then obtained by multiplying the initial confidence level by the decay discount coefficient. The records are aggregated to form a decay record set with depreciation weights. This allows the confidence level of the descriptive items to dynamically reflect the combined depreciation effect of time and maintenance deviations, providing data that is closer to the actual quality status for subsequent vulnerability identification.

[0041] P3. Perform attribute parsing on the attenuation record set to obtain the feature descriptor of the same supply batch.

[0042] In this embodiment of the invention, the step of parsing the attributes of the attenuation record set to obtain the feature descriptor of the same supply batch includes: Extract core attribute fields from the attenuation record set and form a core field group for the same supply batch; The values ​​of related fields within the core field group are merged and aggregated to form aggregated attribute items for the same supply batch; The aggregated attribute items are described and encapsulated to obtain the feature descriptors for the same supply batch.

[0043] When extracting core attribute fields from the attenuation record set, the attenuation record set is matched field by field according to a preset list of core attribute fields. This list lists the field names directly related to material identification and supply characteristics, including material name, specifications, place of origin, mix design number, supply batch number, and quality status description after weighting for loss. During the matching process, fields in the attenuation record set that have the same field name as those in the list, along with their corresponding field values ​​and adjusted confidence levels, are extracted together. These extracted field items are arranged and combined into a field set according to their field names. This field set is the core field group for the same supply batch.

[0044] When merging and aggregating the field values ​​with related relationships within the core field group, the field items within the core field group are grouped according to semantic relevance. When merging the material name and specification model into a material specification identifier, the material name and specification model are used as input, and the standard naming code corresponding to the material name is retrieved from the preset material name standard dictionary. This standard dictionary maintains the naming convention and coding mapping relationship of each material name. If there are multiple variants of the material name in the dictionary, semantic normalization and disambiguation processing is performed. Each variant name is split into a word sequence and matched with the word sequence of each standard name in the dictionary using the longest common subsequence. The standard name code with the highest ratio of common subsequence length to total word number is selected as the output. If there are multiple ratios with the highest values, the standard name code with the highest registered frequency in the dictionary is used. Simultaneously, the specifications and models are decomposed and reorganized according to compliance. Based on the delimiters within the specification and model string, the specifications and models are cut into independent parameter items. Each parameter item is arranged in a fixed order of strength grade, size specification, and material designation to form the initial specification parameter sequence. The strength grade parameter and material designation parameter in the initial specification parameter sequence are checked for mutual exclusivity. If there is a conflict between the material requirement corresponding to the strength grade and the material designation parameter, parameter backtracking correction is triggered. The parameter items related to the conflict in the strength grade parameter and material designation parameter are backtracked to the original specification and model string to re-identify the boundary and replace the conflicting parameter. The initial specification parameter sequence is updated with the parameter items after backtracking correction. Each parameter item in the updated specification parameter sequence is checked against the parameter value range bound to the standard name code in the material name standard dictionary for constraint compliance. Parameter items that exceed the value range are truncated to the upper or lower limit of the value range and a truncation mark is added. The specification parameter sequence that has passed the constraint verification is used as the final specification parameter sequence. A cross-validation of the compatibility between the material name code and the specification parameter sequence is performed. The set of optional specification ranges for the standard name code in the material name standard dictionary is read, and it is checked whether each parameter item in the final specification parameter sequence falls within the corresponding sub-range of the optional specification range set. If all fall within the sub-range, the compatibility verification is deemed successful. The standard name code and the specification parameter sequence are concatenated with a predefined connector, and a compatibility success status flag is added to the front of the result. If any parameter item does not fall within the corresponding sub-range, the compatibility verification is deemed questionable. The standard name code and the specification parameter sequence are concatenated with a predefined connector, and a compatibility questionable status flag is added to the front of the result. At the same time, the names of the parameter items that failed the verification are concatenated at the end of the result to form a traceable anomaly marker. The synthesized string is the material specification identifier.

[0045] When merging the origin identifier and mix proportion number into an origin mix proportion constraint pair, the aggregate gradation range, cementitious material dosage limit, and admixture type requirements for that mix proportion are extracted from the mix proportion filing database using the mix proportion number as the query condition, forming a technical indicator constraint set. Simultaneously, the mining strata information, historical average values ​​of material physicochemical indicators, and annual supply capacity limit for that origin are retrieved from the origin archive using the origin identifier as the query condition, forming an origin capacity file. Each constraint range in the technical indicator constraint set is compared one by one with the corresponding indicator in the origin capacity file. When all corresponding indicators in the origin capacity file fall within the specified range of the technical indicator constraint set, it is considered a successful match. The origin identifier and mix proportion number are then merged in a string concatenation manner, and a successful match status bit is appended to the end of the merged result, forming an origin mix proportion constraint pair.

[0046] When merging all quality status descriptions and their adjusted confidence levels under the same supply batch number into a quality loss profile based on material property type, each quality status description is matched and located layer by layer in the material property classification tree based on its field identifier. The material property classification tree includes three primary categories: mechanical properties, durability properties, and volumetric stability, as well as multiple secondary property types under each primary category. After locating the secondary property type to which the description belongs, the field identifier, field value, and adjusted confidence level of the description are bound into a description tuple. All description tuples belonging to the same secondary property type are grouped together, and the confidence level of the description tuples is determined within the group. The system categorizes descriptive tuples into high-confidence subsets and low-confidence subsets based on the adjusted confidence median of the secondary characteristic type. For the high-confidence subset, the field values ​​of the descriptive tuples are weighted and summed using the adjusted confidence level as the weight to obtain the high-confidence baseline representation value. For the low-confidence subset, the deviation between each field value and the high-confidence baseline representation value is calculated. If the deviation of a low-confidence tuple exceeds the preset deviation tolerance for the secondary characteristic type, the low-confidence tuple is marked as an abnormal deviation tuple and temporarily excluded from weighting; otherwise, it is retained as a fusionable tuple. For tuples with abnormal deviations, the deviation cause is traced. The field values ​​of the tuple with abnormal deviations are compared with the trend consistency of the field value changes of the high-confidence tuple subset within the same group. If the deviation direction of the tuple with abnormal deviations is the same as the change direction of the description tuple corresponding to the most recently collected timestamp within the high-confidence tuple subset, it is determined to be a trend-continuing deviation. The tuple with abnormal deviations is removed from the label and its adjusted confidence is calculated according to the weighting factor before being re-included into the fusionable tuple. If the deviation direction of the tuple with abnormal deviations is opposite to the change direction of the high-confidence tuple subset and the place of origin associated with the tuple with abnormal deviations is different from the place of origin of other description tuples within the same group, it is determined to be a place of origin individuality deviation. The abnormal deviation label is maintained and it is recorded separately as a place of origin anomaly under this secondary characteristic type. The high-confidence subset of tuples is merged with all fusionable tuples, and the field values ​​are weighted and summed using the adjusted confidence level as the weight to obtain the comprehensive loss characterization value for this secondary characteristic type. At the same time, the place-specific items are associated with this secondary characteristic type and then appended to the comprehensive loss characterization value to form an extended characterization structure with specific annotations. This extended characterization structure is the corresponding hierarchical data of this secondary characteristic type in the quality loss profile.

[0047] When describing and encapsulating aggregated attribute items, attribute type labels are added to the material specification identifier, origin and mix proportion constraint pair, and quality loss profile in the aggregated attribute items to indicate the semantic category of each aggregated item. The aggregated items with added labels are assembled into a structured descriptor according to the fixed order of the attribute type labels. The aggregated items in the descriptor are separated by separators. The assembled descriptor fully carries the core information of the batch of materials in the three dimensions of supply source, mix proportion constraint, and quality status. This descriptor is the feature descriptor of the same supply batch.

[0048] The beneficial effect is that by matching each field of the attenuation record set through a pre-set list of core attribute fields, the material name, specifications, origin identification, mix designation number, supply batch number, and quality status description items after loss weighting, along with the adjusted confidence level, are extracted to form a core field group. Key fields directly related to material identification and supply characteristics are accurately extracted from the attenuation record set with loss weights. Fields with semantic relationships within the core field group are merged and aggregated. The material name and specifications are merged into a material specification identification; the origin identification and mix designation number are merged into an origin-mix designation constraint pair; and all quality status description items and their adjusted confidence levels under the same supply batch number are merged according to material characteristic type to form a quality loss profile, generating highly condensed and clearly correlated aggregated attribute items. By attaching attribute type labels to aggregated attribute items and assembling them into structured descriptors in a fixed order, and using separators to distinguish each aggregated item to form feature descriptors, the core information of the same supply batch in the three dimensions of supply source, mix ratio constraints and quality status is uniformly encapsulated into standardized descriptive units. This provides a directly comparable feature carrier for determining place of origin substitution constraints and analyzing mapping intersection relationships in subsequent vulnerability identification.

[0049] P4. Based on the origin substitution constraint of the feature descriptor and the mapping intersection relationship between the current section chainage interval and the mix ratio number, the vulnerability of the attenuation record set is identified, and the dependency label group of the attenuation record set is obtained. In this embodiment of the invention, the mapping intersection relationship between the origin substitution constraint based on feature descriptors and the current section chainage interval and mix proportion number is used to identify the vulnerability of the attenuation record set, resulting in a dependency tag group for the attenuation record set, including: Constraint fields are extracted from the feature descriptors to obtain the origin identifier and blending ratio number for the same supply batch; Based on the mapping and intersection relationship between the current section chainage interval and the mix proportion number, determine whether there is a compliant alternative origin other than the origin identifier for the mix proportion number; When it is determined that there is no compliant alternative origin, the origin identifier is marked as an irreplaceable origin and merged into the dependency tag group of the decay record set.

[0050] Based on the mapping intersection between the current lot number interval and the mix designation number, determine whether the mix designation number has a compliant alternative origin other than the origin identifier, including: Query the mix proportion for the current chainage interval to obtain the mix proportion number corresponding to the current chainage interval; By binding the corresponding blending ratio number to the place of origin for traceability, a set of compliant places of origin for the corresponding blending ratio number is obtained; By comparing the differences in the compliant origin set, if the compliant origin set only contains the origin identifier, it is confirmed that there is no compliant alternative origin.

[0051] When extracting constraint fields from feature descriptors, locate the material specification identifier, origin mix ratio constraint pair, and quality loss profile with attribute type labels from the feature descriptors. Read the origin identifier field value and mix ratio number field value from the origin mix ratio constraint pair respectively, and use the read origin identifier and mix ratio number as the origin identifier and mix ratio number of the same supply batch.

[0052] When querying the mix proportion for the current section's chainage interval, the current construction section's chainage interval is used as the query basis in the section's mix proportion mapping table of the engineering design. The mix proportion number record associated with the chainage interval is searched, and the unique mix proportion number corresponding to that interval is matched. This mix proportion number is the mix proportion number corresponding to the current section's chainage interval.

[0053] When binding the corresponding mix proportion number to trace the place of origin, the corresponding mix proportion number is used as the search condition in the compliant place of origin registration database of engineering materials. All place of origin records that have been registered and approved under the mix proportion number are searched. The place of origin name in each of the retrieved place of origin records is extracted one by one to form a place of origin list. This place of origin list is the compliant place of origin set of the corresponding mix proportion number.

[0054] When performing a difference comparison on the compliant origin set, each origin name in the compliant origin set is compared one by one with the origin identifier of the same supply batch to determine whether there are other origin names in the compliant origin set that are different from the origin identifier. If all origin names in the compliant origin set are completely consistent with the origin identifier and there are no other origin names, it is determined that there is no compliant alternative origin other than the origin identifier for the blending ratio number. The origin identifier is then marked as an irreplaceable origin and added to the dependency tag group of the attenuation record set.

[0055] The beneficial effects are as follows: By extracting constraint fields from feature descriptors, the origin identifier and mix proportion number are accurately separated from the descriptor body, restoring the encapsulated feature information into independent constraint conditions that can be used for logical judgment. Based on the current section chainage interval, a unique corresponding mix proportion number is obtained by querying the section mix proportion mapping table in the engineering design, establishing a precise correspondence between the construction section and the material technical requirements. Using this mix proportion number as a search condition, origin traceability is performed in the engineering material compliant origin registration database, extracting all registered and approved origin records to form a compliant origin set, comprehensively gathering legal supply sources that meet the technical standards of this mix proportion. The compliant origin set is compared one by one with the origin identifiers of the same supply batch. When the compliant origin set contains only this origin identifier and no other origin name, it is determined that there is no compliant alternative origin. This origin identifier is then marked as an irreplaceable origin and added to the dependency marker group, achieving automatic identification and precise labeling of the unique dependent origin in the supply chain. This identifies key vulnerable nodes that need to be prioritized for protection in subsequent backtracking and optimization to construct an alternative topology map.

[0056] P5. Using the non-substitutable production sites marked in the dependency tag group as anchor points, backtrack and select the best historical supply records to obtain a substitution topology map oriented towards non-substitutable production sites. In this embodiment of the invention, the step of using the non-substitutable production sites marked in the dependency tag group as anchor points to backtrack and optimize historical supply records to obtain a substitution topology map oriented towards non-substitutable production sites includes: By performing correlation filtering on historical supply records, a subset of related records with supply relationships to irreplaceable production areas is obtained; Perform blending ratio matching on a subset of associated records to obtain alternative supply source records that meet the blending ratio number corresponding to the irreplaceable production area; The records of the candidate supply sources are sorted by impedance to obtain a hierarchy of merits for each candidate supply source. Based on the hierarchy of superiority and inferiority, a topology is constructed for the reachable paths between alternative supply source records and irreplaceable production sites to obtain an alternative topology map oriented towards irreplaceable production sites.

[0057] When performing correlation filtering on historical supply records, the irreplaceable production areas marked in the dependency tag group are used as the filtering conditions. All supply record entries are retrieved in the historical supply record database. The production area field value of each supply record is read one by one and compared with the irreplaceable production area. Records whose production area field value is completely consistent with the name of the irreplaceable production area are retained. At the same time, the historical supply chain upstream and downstream of each retained record is traced to extract all related records that have had a peer supply relationship, transshipment relationship or distribution relationship with the irreplaceable production area. These related records are aggregated into a record subset to obtain a subset of related records that have a supply relationship with the irreplaceable production area.

[0058] When performing mix proportion matching on a subset of associated records, the mix proportion number corresponding to the non-substitutable origin is read from the feature descriptor corresponding to the dependency tag group. This mix proportion number is used as the matching benchmark. Each record in the subset of associated records is traversed, and the mix proportion number field value marked on each record is extracted. The extracted mix proportion number field value is compared with the matching benchmark for consistency. Only records whose mix proportion number field value is completely consistent with the matching benchmark are retained. The retained records constitute the alternative supply source records that meet the mix proportion number corresponding to the non-substitutable origin.

[0059] When sorting the candidate supply source records by impedance, the delivery time data, transportation distance data, historical performance completion rate, and quality pass rate of each candidate supply source are extracted from the records. The delivery time is sorted in ascending order and mapped to a timeliness score, the transportation distance is sorted in ascending order and mapped to a distance score, the historical performance completion rate is sorted in descending order and mapped to a performance score, and the quality pass rate is sorted in descending order and mapped to a quality score. The timeliness score, distance score, performance score, and quality score of each candidate supply source are weighted and summed to obtain a comprehensive impedance score. All candidate supply sources are sorted from low to high according to the comprehensive impedance score. The lower the comprehensive impedance score, the smaller the impedance of the supply substitution and the higher the priority. The sorting result is the superiority and inferiority sequence of each candidate supply source.

[0060] When constructing the topology of reachable paths between candidate supply source records and irreplaceable production locations based on the priority-priority hierarchy, the irreplaceable production location is taken as the root node of the topology graph. Candidate supply sources are sequentially selected from the priority-priority hierarchy as candidate nodes. For each candidate node, all actual transportation path records that have occurred between that candidate node and the irreplaceable production location are extracted from the associated record subset. Each transportation path record includes the sequence of transit nodes, the transportation mode between adjacent nodes, the average dwell time of each transit node, and the standard transportation timeliness between adjacent nodes. The transportation mode between all adjacent nodes on the transportation path is determined as the type of directed edge connecting that segment. If the transportation mode is road transportation, the directed edge is assigned road timeliness and road capacity attributes; if the transportation mode is rail transportation, the directed edge is assigned rail timeliness and rail capacity attributes; if the transportation mode is waterway transportation, the directed edge is assigned waterway timeliness and waterway capacity attributes. Nodes are added to the topology graph in the order they appear in the transportation path, and the average dwell time of each node is attached as a node weight. All intermediate nodes are expanded layer by layer between the root node of the irreplaceable production area and each candidate node. Adjacent nodes are connected by directed edges, and the directed edge type, timeliness attribute, and transportation capacity attribute are written into the directed edge until a complete directed path is formed from the root node through each intermediate node to the candidate node. When there are multiple historical transportation paths for the same candidate node, each path is retained and expanded into an independent path. At the same time, the quality stability rating and current capacity saturation of each candidate node are extracted from the candidate supply source record as candidate node attributes and attached to the corresponding candidate node. The root node of the irreplaceable production area, all intermediate nodes, and all candidate nodes are completely connected with their respective attributes to form a multi-path directed network. This multi-path directed network is the alternative topology graph for the irreplaceable production area.

[0061] The beneficial effect is that historical supply records are correlated and filtered using the irreplaceable production locations marked in the dependency tag group as anchors. Through comparison of production location field values ​​and tracing the upstream and downstream of the supply chain, all records that have had supply, transshipment, and distribution relationships with irreplaceable production locations are extracted to form a subset of associated records, expanding the scope of backtracking from isolated production location records to a complete supply relationship network. The mix proportion number is read from the feature descriptor as a matching benchmark to perform consistency comparison on the subset of associated records. Only records whose mix proportion number is completely consistent with the mix proportion number corresponding to the irreplaceable production location are retained as candidate supply source records, ensuring that each candidate supply source matches the technical requirements of the current project section. Delivery timeliness, transportation distance, historical performance completion rate, and quality pass rate are extracted from the candidate supply source records and subjected to multi-dimensional scoring mapping and equal weighting to obtain a comprehensive impedance score. These scores are arranged from low to high to form a hierarchy of superiority and inferiority; the lower the comprehensive impedance score, the smaller the impedance of supply substitution and the higher the priority. Based on the hierarchy of superior and inferior products, with the irreplaceable production location as the root node, a directed pathway network is formed by expanding the transit nodes and logistics connection methods of alternative supply sources layer by layer, thus obtaining an alternative topology map. This provides a clear overview of alternative paths for subsequent strategy matching and output of supply guarantee instruction sets.

[0062] P6. Perform strategy matching on the alternative topology map to obtain a supply guarantee instruction set for irreplaceable production areas.

[0063] In this embodiment of the invention, the step of performing strategy matching on the alternative topology map to obtain a supply guarantee instruction set for irreplaceable production locations includes: Structural analysis is performed on the alternative topology graph to obtain its topological characteristic parameters; Threshold comparison of topological feature parameters yields risk level identifiers for alternative topological maps; By mapping risk level identifiers to instructions, a set of supply guarantee instructions for irreplaceable production areas is obtained.

[0064] When performing structural analysis on the alternative topology graph, all directed paths from the non-substitutable origin root node to each candidate supply source node are traversed. For each directed path, the node weights of all intermediate nodes are extracted and accumulated to obtain the cumulative transit dwell time. The timeliness attributes of all directed edges on the same path are extracted and accumulated to obtain the cumulative transportation time. The cumulative transit dwell time and the cumulative transportation time are added together to obtain the total link equivalent time of the path. The total link equivalent time is used as the path depth of the path. The path depths of all paths are aggregated to obtain the path depth set. The total number of candidate nodes in the alternative topology graph is counted as the number of candidate supply sources. If the quality stability rating attached to a candidate node is lower than the preset quality stability lower limit, the candidate node is not included in the number of candidate supply sources but is separately included in the set of downgraded supply sources. The total number of directed edges in the alternative topology graph is used as the edge size. Directed edges are further divided according to transportation mode: the number of directed edges for highways, railways, and waterways. The distribution of the number of edges for each transportation mode is used as the edge size. It is checked whether there are candidate nodes that are unreachable from the root node. If so, these candidate nodes are listed as isolated candidate nodes, and the number of isolated candidate nodes is recorded as the isolated node count. Simultaneously, the reachability of candidate nodes in the degraded supply source set to the root node is also checked; if they are unreachable, they are also included in the isolated node count. The resulting path depth set, number of candidate supply sources, edge size, and number of isolated nodes together constitute the topological characteristic parameters of the alternative topology graph.

[0065] When performing threshold comparisons on topological feature parameters, the number of candidate supply sources in the topological feature parameters is compared with a preset lower limit threshold for the number of supply sources, the size of connected edges is compared with a preset lower limit threshold for the size of connected edges, the number of isolated nodes is compared with a preset upper limit threshold for isolated nodes, and the maximum path depth in the path depth set is compared with a preset upper limit threshold for path depth. If the number of candidate supply sources is lower than its lower limit threshold, the size of connected edges is lower than its lower limit threshold, the number of isolated nodes is higher than its upper limit threshold, or the maximum path depth is higher than its upper limit threshold, then the alternative topological graph is determined to be of a high-risk level and assigned a high-risk level label. If none of the four comparison results trigger a high-risk determination, then the alternative topological graph is determined to be of a low-risk level and assigned a low-risk level label.

[0066] When mapping instructions to risk level identifiers, the instruction set configuration scheme corresponding to the risk level identifier is retrieved from the preset supply security instruction mapping library. The supply security instruction mapping library stores the mapping relationship from high-risk level identifiers to high-risk instruction sets and from low-risk level identifiers to low-risk instruction sets. The high-risk instruction set includes instructions to initiate multi-source parallel procurement, instructions to activate emergency reserve release, and instructions to increase transportation priority. The low-risk instruction set includes instructions to maintain regular procurement, instructions to maintain existing inventory rotation, and instructions to periodically monitor supply status. All instructions in the retrieved instruction set configuration scheme are extracted one by one and arranged in priority order. The arranged instruction sequence is the supply security instruction set for irreplaceable production areas.

[0067] The beneficial effects are as follows: By performing structural analysis on the alternative topology graph, traversing all directed paths, and statistically analyzing path depth, the number of alternative supply sources, the size of connecting edges, and the number of isolated nodes to construct topological feature parameters, the availability, connectivity, and reachability of alternative paths are transformed into quantifiable evaluation indicators. Multiple preset thresholds are used to compare the topological feature parameters one by one. Based on whether the number of alternative supply sources is sufficient, the size of connecting edges is adequate, the number of isolated nodes exceeds the limit, and the maximum path depth is excessive, the risk level of the alternative topology graph is automatically determined and assigned a corresponding label, achieving standardized hierarchical evaluation of the feasibility of supply substitution for irreplaceable production areas. Based on the risk level label, matching instruction set configuration schemes are retrieved from the supply guarantee instruction mapping library. High-risk levels trigger multi-source parallel procurement instructions, emergency reserve release instructions, and instructions to increase transportation priority; low-risk levels trigger instructions to maintain regular procurement, maintain existing inventory rotation, and periodically monitor supply status. Through hierarchical mapping, differentiated and orderly supply guarantee instruction sets are output, ensuring that guarantee measures for irreplaceable production areas are accurately matched with the degree of topological risk, improving the targeting and execution efficiency of response measures when key supply nodes are constrained.

[0068] like Figure 2 The diagram shown is a functional block diagram of an engineering materials supply chain risk early warning system based on multi-source data fusion provided in an embodiment of the present invention, used for execution. Figure 1 The methods described herein are used to achieve the corresponding technical effects.

[0069] The engineering materials supply chain risk early warning system 100 based on multi-source data fusion described in this invention can be installed in an electronic device. Depending on the functions implemented, the engineering materials supply chain risk early warning system 100 may include a semantic parsing module 101, a loss weighting module 102, an attribute parsing module 103, a vulnerability identification module 104, a backtracking optimization module 105, and a strategy matching module 106. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0070] In this embodiment, the functions of each module / unit are as follows: The semantic parsing module 101 is used to perform multi-dimensional semantic parsing on engineering supply association information from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information. The loss weighting module 102 is used to perform loss weighting on the standard record set corresponding to the same supply batch based on the material aging characteristics and maintenance deviation implied in the standard record set, so as to obtain the attenuation record set of the same supply batch. The attribute parsing module 103 is used to perform attribute parsing on the attenuation record set to obtain the feature descriptor of the same supply batch. The vulnerability identification module 104 is used to identify the vulnerability of the attenuation record set based on the origin substitution constraint of the feature descriptor and the mapping intersection relationship between the current section chainage interval and the mix ratio number, and obtain the dependency tag group of the attenuation record set. The backtracking optimization module 105 is used to backtrack and optimize historical supply records using the irreplaceable production sites marked in the dependency tag group as anchor points, so as to obtain an alternative topology map oriented towards irreplaceable production sites. The strategy matching module 106 is used to perform strategy matching on the alternative topology map to obtain a supply guarantee instruction set for irreplaceable production areas.

[0071] Figure 3 The comparison shows the frequency of different supply guarantee instructions triggered for high-risk and low-risk batches within a unit of time. Instructions for high-risk batches, primarily "multi-source procurement," "emergency reserves," and "increased transportation priority," were triggered more frequently. Instructions for low-risk batches, mainly "routine procurement," "inventory rotation," and "status monitoring," were triggered significantly more frequently than those for high-risk batches. This clearly demonstrates the actual effectiveness of the system in outputting refined and differentiated guarantee strategies based on risk level differences.

[0072] Figure 4Using topological characteristic parameters as the analysis object, this paper demonstrates the boundary relationship between key indicators such as the number of alternative supply sources, path depth, and number of connected edges and risk levels when determining the risk level of alternative topology graphs. Curves clearly identify the risk assessment boundaries under different parameter values. For example, when the number of alternative supply sources is below the lower threshold, the path depth exceeds the upper threshold, or the size of connected edges is too small, the topology graph will be classified as high-risk; conversely, it will be classified as low-risk. This provides a quantitative classification basis for subsequent command mapping and supply security decisions.

[0073] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0074] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0075] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0076] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0077] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for early warning of risks in the engineering materials supply chain based on multi-source data fusion, characterized in that, The method includes: P1. Perform multi-dimensional semantic parsing on engineering supply association information originating from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information; P2. Based on the material aging characteristics and maintenance deviations implicit in the standard record set, the standard record set corresponding to the same supply batch is weighted for depreciation to obtain the depreciation record set of the same supply batch. P3. Perform attribute parsing on the attenuation record set to obtain the feature descriptor of the same supply batch; P4. Based on the origin substitution constraint of the feature descriptor and the mapping intersection relationship between the current section chainage interval and the mix ratio number, the vulnerability of the attenuation record set is identified to obtain the dependency tag group of the attenuation record set. P5. Using the non-substitutable production areas marked in the dependent tag group as anchor points, backtrack and select the best historical supply records to obtain a substitution topology map oriented towards the non-substitutable production areas. P6. Perform strategy matching on the alternative topology map to obtain a supply guarantee instruction set for the irreplaceable production location.

2. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 1, characterized in that, The process of performing multi-dimensional semantic parsing on engineering supply association information originating from multiple heterogeneous data interfaces yields a standardized record set of engineering supply association information, including: Field decomposition is performed on engineering supply association information originating from multiple heterogeneous data interfaces to obtain the original field set of the engineering supply association information; Based on semantic mapping rules, the original field set is transformed to obtain the normalized field set of the original field set; The normalized field set is reconstructed to obtain a standardized record set of the engineering supply-related information.

3. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 1, characterized in that, The method of assigning depreciation weights to the standard record set corresponding to the same supply batch based on the material aging characteristics and maintenance deviations implicit in the standard record set, to obtain the depreciation record set of the same supply batch, includes: Time effect analysis is performed on the standard record set to obtain the set of attenuation parameters for the same supply batch; The monitoring temperature and humidity records of the same supply batch during transportation and storage are obtained to obtain the environmental measurement data of the same supply batch; Based on the set of attenuation parameters and the measured environmental data, the descriptive items reflecting the material quality status in the standard record set are weighted according to their depreciation, thus obtaining the attenuation record set for the same supply batch.

4. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 3, characterized in that, The confidence level of the descriptive items reflecting the material quality status in the standard record set is adjusted based on the attenuation parameter set and the environmental measured data to obtain the attenuation record set for the same supply batch, including: The attenuation parameter set is subjected to parameter extraction to obtain the attenuation rate of the same supply batch; Accelerated extrapolation is performed on the measured environmental data to obtain the environmental acceleration factor for the same supply batch; The duration of each descriptive item that reflects the material quality status in the standard record is obtained by acquiring the duration of each descriptive item. Initial confidence levels are extracted for each descriptive term to obtain the initial confidence level of each descriptive term. Then, confidence discounting is performed on each descriptive term according to the following formula to obtain the adjusted confidence level of each descriptive term: ； In the formula, As the reference decay rate, As an environmental accelerator, The duration is... As the initial confidence level, To adjust the confidence level; The adjusted confidence levels are aggregated to obtain a set of attenuation records for the same supply batch.

5. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 1, characterized in that, The step of parsing the attributes of the attenuation record set to obtain feature descriptors for the same supply batch includes: Extract core attribute fields from the attenuation record set to form a core field group for the same supply batch; The values ​​of fields with related relationships within the core field group are merged and aggregated to form aggregated attribute items for the same supply batch; The aggregated attribute items are described and encapsulated to obtain the feature descriptor of the same supply batch.

6. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 1, characterized in that, The vulnerability identification of the attenuation record set is performed based on the mapping intersection relationship between the origin substitution constraint based on the feature descriptor and the current section chainage interval and mix proportion number, resulting in a dependency tag group for the attenuation record set, including: Constraint fields are extracted from the feature descriptor to obtain the origin identifier and blending ratio number of the same supply batch; Based on the mapping intersection between the current section chainage interval and the mix proportion number, it is determined whether the mix proportion number has a compliant alternative origin other than the origin identifier; When it is determined that there is no compliant alternative origin, the origin identifier is marked as an irreplaceable origin and added to the dependency tag group of the decay record set.

7. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 6, characterized in that, The determination of whether a mix designation number exists as a compliant alternative to the origin identifier, based on the mapping intersection between the current lot number interval and the mix designation number, includes: Perform a mix design ratio lookup for the current chainage interval to obtain the mix design ratio number corresponding to the current chainage interval. By binding the corresponding blending ratio number to the place of origin for traceability, a set of compliant places of origin for the corresponding blending ratio number is obtained; The compliant origin set is compared for differences. If the compliant origin set contains only the origin identifier, it is confirmed that there is no compliant alternative origin.

8. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 1, characterized in that, The step of using the non-substitutable production sites marked in the dependency tag group as anchor points to backtrack and optimize historical supply records to obtain a substitution topology map oriented towards the non-substitutable production sites includes: By performing correlation filtering on historical supply records, a subset of related records with supply relationships to the aforementioned irreplaceable production areas is obtained; The associated record subset is matched with the blending ratio to obtain alternative supply source records that meet the blending ratio number corresponding to the irreplaceable production area; The candidate supply source records are sorted by impedance to obtain a hierarchy of merits for each candidate supply source. Based on the superiority-inferiority hierarchy, a topology is constructed for the reachable paths between the candidate supply source records and the irreplaceable production sites to obtain an alternative topology map oriented towards the irreplaceable production sites.

9. The method for risk early warning of engineering materials supply chain based on multi-source data fusion as described in claim 1, characterized in that, The step of performing strategy matching on the alternative topology map to obtain a supply guarantee instruction set for the irreplaceable production areas includes: The alternative topology graph is subjected to structural analysis to obtain the topological feature parameters of the alternative topology graph; Threshold comparison is performed on the topological feature parameters to obtain the risk level identifier of the alternative topological map; By mapping the risk level identifier to an instruction set, a supply guarantee instruction set for the irreplaceable production area is obtained.

10. A risk early warning system for engineering materials supply chain based on multi-source data fusion, used to implement the risk early warning method for engineering materials supply chain based on multi-source data fusion as described in any one of claims 1-9, the system comprising: The semantic parsing module is used to perform multi-dimensional semantic parsing on engineering supply association information from multiple heterogeneous data interfaces to obtain a standardized record set of engineering supply association information; The loss weighting module is used to assign loss weights to the standard record set corresponding to the same supply batch based on the material aging characteristics and maintenance deviations implicit in the standard record set, so as to obtain the attenuation record set of the same supply batch. The attribute parsing module is used to parse the attributes of the attenuation record set to obtain the feature descriptors of the same supply batch. The vulnerability identification module is used to identify the vulnerability of the attenuation record set based on the origin substitution constraint of the feature descriptor and the mapping intersection relationship between the current section chainage interval and the mix ratio number, and to obtain the dependency tag group of the attenuation record set. The backtracking optimization module is used to backtrack and optimize historical supply records using the irreplaceable production areas marked in the dependency tag group as anchor points, so as to obtain an alternative topology map oriented towards the irreplaceable production areas. The strategy matching module is used to perform strategy matching on the alternative topology map to obtain a supply guarantee instruction set for the irreplaceable production location.

Images