An electrical transmission equipment sales management method and system based on a knowledge graph

By adopting a knowledge graph-based sales management method for electrical drive equipment, the matching status of equipment across multiple constraint dimensions is obtained and verified. This solves the problem that existing technologies cannot simultaneously meet complex constraints, thus achieving accuracy and feasibility in equipment selection.

CN123117724APending Publication Date: 2026-11-17JIANGSU MENGLIYUAN SOFTWARE TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611600873.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-16
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the sales management of electrical drive equipment, existing technologies simplify multiple condition combinations into single-condition matching, resulting in recommended equipment failing to simultaneously meet complex constraints, thus affecting the feasibility of equipment selection.

Method used

By using a knowledge graph-based approach, multiple constraint dimensions of electrical drive equipment are obtained, a unified association identifier is configured, candidate targets are retrieved and local matching status is recorded, and synchronous verification records are generated to ensure that the equipment meets the standards in multiple dimensions simultaneously.

Benefits of technology

It enables joint verification of multi-dimensional constraints, ensuring that recommended equipment meets standards in multiple dimensions such as power, environment, and certification, thereby improving the accuracy of equipment selection and sales conversion efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117724A_ABST
    Figure CN123117724A_ABST
Patent Text Reader

Abstract

The application relates to the technical field of electrical transmission equipment sales management, and discloses an electrical transmission equipment sales management method and system based on a knowledge graph, which comprises the following steps: obtaining multiple constraint dimension demands of electrical transmission equipment, and configuring a unified correlation identifier for each constraint dimension demand; based on a preset correlation relationship model, a candidate target matched with each constraint dimension demand is searched respectively, and a local matching state between each candidate target and each constraint dimension demand is recorded; according to the correlation identifier, each candidate target is allocated with corresponding local matching states, so as to generate a synchronous check record; based on the synchronous check record, a preset matching state and a preset performance threshold, a compliance recommendation object is determined from the candidate target. The application significantly improves the technical executability and scheme accuracy of the sales selection result.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of sales management technology for electrical drive equipment, and more specifically, to a knowledge graph-based method and system for sales management of electrical drive equipment. Background Technology

[0002] In existing sales management of electrical drive equipment, knowledge graph technology is typically used to semantically represent and reason about business entities such as equipment, customers, operating conditions, and certifications, as well as their relationships. During the selection process for non-standard projects, after the sales engineer enters the project requirements, the system matches and sorts them based on the static relationships between equipment nodes and power ranges, environmental conditions, and certification conditions, and outputs recommended models. This approach expresses various constraints as independent relationships between equipment nodes and performs reasoning along the relationship paths, effectively filtering equipment when handling single or simple condition selections.

[0003] However, in actual non-standard project selection, customer needs often involve complex combinations of conditions such as high-temperature, high-humidity, and corrosive environments, specific heavy-load power requirements, and industry access certifications. Existing technologies decompose these conditions into three independent static relationships: power segment relationships, environmental condition relationships, and certification relationships. The inference chain only hits the device node along a single relationship. Because knowledge graphs treat these conditions as independent static edges, hitting any edge during matching can increase the recommendation confidence of the device or directly reach the device node, lacking joint verification of whether the same device node simultaneously meets all constraints. This results in output devices often only meeting some conditions, failing subsequent technical reviews and certification audits, severely impacting the executability of sales selection recommendations and forcing sales engineers to perform extensive manual verification.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this application discloses a knowledge graph-based sales management method and system for electrical drive equipment. The aim is to solve the technical problem in existing technologies where multiple conditions are simplified into single-condition matching, resulting in recommended equipment failing to simultaneously meet complex constraints, thus affecting the feasibility of equipment selection.

[0006] The technical solution of this application is as follows: In a first aspect, this application discloses a sales management method for electrical drive equipment based on knowledge graphs, the steps of which include: Obtain multiple constraint dimensions of requirements for electrical drive equipment and configure a unified association identifier for each constraint dimension requirement; Based on the pre-defined association model, candidate targets that match the requirements of each constraint dimension are retrieved, and the local matching status between each candidate target and the requirements of each constraint dimension is recorded. Based on the association identifier, assign corresponding local matching states to each candidate target to generate a synchronization verification record; Based on synchronous verification records, preset matching status, and preset performance thresholds, compliant recommended objects are determined from candidate targets.

[0007] This technical solution enables joint verification of multi-dimensional constraints, ensuring that recommended devices meet standards simultaneously in multiple dimensions such as power, environment, and certification, thus fundamentally solving the logical loopholes caused by single-path reasoning.

[0008] Optionally, the relationship model is a knowledge graph, whose nodes include equipment model nodes and performance nodes. The performance nodes include power segment nodes, environmental protection level nodes, corrosive environment adaptation nodes, and industry certification nodes. The nodes are connected by static relationship edges, which carry performance indicator attribute values, including rated power values, overload capacity percentage, and IP protection level parameters. Based on a pre-defined association model, candidate targets that match the requirements of each constraint dimension are retrieved, and the local matching status between each candidate target and the requirements of each constraint dimension is recorded, including: Starting from the performance nodes corresponding to the requirements of each constraint dimension, we traverse and reason along the static relationship edges in the relationship model to the device model nodes to determine the candidate objects. Record the static relation edges that the association model passes through during the process of determining candidate objects, and use the performance index attribute values ​​in the static relation edges as local matching states.

[0009] Optionally, the step of assigning corresponding local matching states to each candidate target based on the association identifier to generate a synchronization verification record includes: Based on the association identifier, a multi-dimensional state matrix is ​​created for each candidate target, and the dimensions of the multi-dimensional state matrix correspond one-to-one with the multiple constraint dimensions requirements; The retrieved local matching states are filled into the corresponding dimensions of the multidimensional state matrix to form a synchronous verification record of the candidate target for the electrical drive equipment.

[0010] Optionally, after determining compliant recommendation targets from candidate targets, the method further includes: For candidate targets that have not been identified as compliant recommendation objects, retrieve dimension items with matching status of "not met" or "not matched" from the synchronous verification records; The dimension item is identified as the missing constraint dimension requirement, and based on the preset business rule table and preset substitutability rules, it is determined whether the missing constraint dimension requirement belongs to a substitutable constraint that can be made up by external means. If the missing constraint dimension requirement is a replaceable constraint, then a compensation scheme suggestion is generated for the candidate object; If the missing constraint dimension requirement is not a replaceable constraint, then a permanent block is applied to the candidate object.

[0011] Optionally, after generating compensation scheme suggestions for candidate objects, the method further includes: Obtain the secondary impact correction coefficient corresponding to the compensation plan suggestion. The secondary impact correction coefficient is used to characterize the negative physical impact of the compensation accessories on the performance of the equipment itself. The original performance index of the candidate target is weighted and calculated based on the secondary impact correction coefficient to obtain the corrected performance index. The corrected performance index is equal to the product of the original performance index and the secondary impact correction coefficient. The revised performance metrics are compared with the performance thresholds a second time. If the corrected performance metrics exceed the performance threshold, the original compensation plan should be maintained. If the corrected performance metrics fall below the performance threshold, the proposed compensation scheme will be withdrawn, and the corresponding candidate will be permanently blocked.

[0012] Optionally, after determining compliant recommendation targets from candidate targets, the method further includes: Based on the judgment results, priority tiers are divided, with compliant recommended objects classified as the first priority tier and candidate targets with alternative constraints classified as the second priority tier. Extract the performance metrics of the compliant recommended objects under the requirements of various constraint dimensions, and calculate the redundancy coefficient of the performance metrics relative to the preset bottom line requirements; Based on the redundancy coefficient and the preset historical sales conversion rate, calculate the comprehensive recommendation score of the compliant recommended objects, and sort the objects in the first priority tier in descending order according to the comprehensive recommendation score; Perform cost penalty deduction on candidate targets in the second priority tier and sort them in descending order after the ranking results of the first priority tier; Based on the order of the first priority tier and the second priority tier, generate and output the recommended ranking results.

[0013] Optionally, the steps for calculating the comprehensive recommendation score of compliant recommended objects based on the redundancy coefficient and the preset historical sales conversion rate include: Obtain historical maintenance records and component failure rates of compliant recommended objects, and combine them with preset equipment deployment environment parameters to determine the performance degradation prediction curve of compliant recommended objects; Based on the performance degradation prediction curve, calculate the performance envelope area of ​​the compliant recommended object within the preset service life. The redundancy coefficient is corrected based on the area of ​​the performance envelope to obtain the time redundancy score, which is the product of the area of ​​the performance envelope and the redundancy coefficient. The timeliness redundancy score is weighted with the historical sales conversion rate to obtain a comprehensive recommendation score for compliant recommended targets.

[0014] Optionally, the method further includes: When none of the candidate targets are identified as compliant recommendations, all candidate targets subject to permanent blocking will be diverted to the manual review queue. Cluster analysis is performed on the reasons for the missing candidate targets in the manual review queue to generate an early warning report; Obtain feedback from the manual review queue, and update the entity relationships in the association model based on the feedback, or adjust the parsing rule parameters of the constraint dimension requirements.

[0015] Optionally, before determining compliant recommendation objects from candidate targets based on synchronous verification records, preset matching states, and preset performance thresholds, the method further includes: Identify dimension pairs that are physically or logically coupled among multiple constraint dimensions and use them as a collaborative verification dimension group; The interaction interference coefficients of candidate targets under the collaborative verification dimension group are extracted from the correlation model. The interaction interference coefficients are used to characterize the performance loss when different dimensional requirements are implemented simultaneously on the same hardware carrier. Based on the cross-interference coefficient, the corresponding local matching state in the synchronous verification record is reduced to generate a collaborative correction record; Based on synchronous verification records, preset matching status, and preset performance thresholds, compliant recommended targets are determined from candidate targets, including: Candidate targets are jointly determined based on synchronous verification records and collaborative correction records. A candidate target is determined to be a compliant recommendation object if and only if the corrected state of the candidate target after considering the interference between dimensions is still in a matching state and meets the performance threshold.

[0016] Secondly, this application also discloses a knowledge graph-based sales management system for electrical drive equipment, used to execute the steps in the above method, characterized in that it includes: The requirement acquisition module is used to acquire multiple constraint dimensions of requirements for electrical drive equipment and configure a unified association identifier for each constraint dimension requirement; The retrieval record module is used to retrieve candidate targets that match the requirements of each constraint dimension based on a preset association model, and record the local matching status between each candidate target and the requirements of each constraint dimension. The record building module is used to assign corresponding local matching states to each candidate target based on the association identifier in order to generate synchronous verification records; The joint judgment module is used to determine compliant recommended objects from candidate targets based on synchronous verification records, preset matching status, and preset performance thresholds.

[0017] In summary, this application provides a knowledge graph-based sales management method and system for electrical drive equipment. By constructing a knowledge graph-based synchronous verification mechanism, it unifies and matrixes the multi-dimensional constraints in the selection of electrical drive equipment, effectively solving the logical loopholes caused by independent reasoning in existing technologies. By introducing collaborative verification dimension groups and interactive interference coefficients, this application can accurately assess performance losses under multiple condition combinations, ensuring the true compliance of recommended equipment under complex operating conditions, and significantly improving the accuracy of non-standard project selection and sales conversion efficiency. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a knowledge graph-based sales management method for electrical drive equipment, provided as an embodiment of this application.

[0019] Figure 2 This is a schematic diagram of the structure of a knowledge graph-based sales management system for electrical drive equipment, provided as an embodiment of this application.

[0020] Figure 3 This is a schematic diagram of the non-compliance handling process provided in the embodiments of this application.

[0021] Labeling Explanation: 210, Requirement Acquisition Module; 220, Search Record Module; 230, Record Construction Module; 240, Joint Decision Module. Detailed Implementation

[0022] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0023] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] In the technical application scenarios of sales management of electrical drive equipment, when facing customized selection for non-standard projects, customers' operating conditions often present complex characteristics with multiple intertwined physical conditions. When traditional solutions screen equipment through semantic relationship networks, they usually decouple multi-dimensional requirements such as power, environment, and certification into discrete independent conditions. The reasoning mechanism triggers equipment recall and confidence accumulation based solely on the connectivity of a single relationship path, easily overlooking the strong constraint that multiple stringent conditions must be simultaneously satisfied on the same physical entity of the equipment. This results in the system recommending equipment models that may only meet the requirements of a single dimension but have deficiencies in other rigid dimensions, making them unsuitable for direct application in engineering technical solutions.

[0025] In this regard, firstly, referring to Figure 1 and Figure 3 This application proposes a sales management method for electrical drive equipment based on knowledge graphs. The steps of this method include: S1. Obtain multiple constraint dimension requirements of the electrical drive equipment and configure a unified association identifier for each constraint dimension requirement; S2. Based on the preset association model, retrieve candidate targets that match the requirements of each constraint dimension, and record the local matching status between each candidate target and the requirements of each constraint dimension. S3. Based on the association identifier, assign corresponding local matching states to each candidate target to generate a synchronization verification record; S4. Based on synchronous verification records, preset matching status, and preset performance thresholds, determine compliant recommended objects from candidate targets.

[0026] Constraint requirements refer to the specific indicators and qualification restrictions proposed by customers based on their actual application conditions during the selection of non-standard projects. These include electrical power range, physical protection capabilities of the working environment, corrosion resistance, and specific industry access permits. Constraint requirements can originate from structured fields entered in the sales management system, or from semi-structured descriptions in project description texts, inquiry form attachments, or historical communication minutes.

[0027] The association identifier refers to a globally unified association mark configured by the system for multiple constraint dimensions in a single non-standard selection task. It is used to indicate that various constraints belong to the same project task and need to be established collaboratively on the same physical device. This association identifier is preferably obtained using a globally unique task identifier generation mechanism. It can be formed by combining the project number, timestamp fragment, and serial number, or it can directly call the existing project primary key in the sales management system as a unified identifier.

[0028] Local matching status refers to the degree of satisfaction or related attribute data presented by a candidate device when facing a single constraint dimension requirement, as revealed through data retrieval. Local matching status includes not only status markers such as "hit" or "miss," but also the specific attribute content supporting that status.

[0029] Synchronization verification records refer to a global verification data set that aggregates the local matching states of a candidate device under all constraint dimensions and is constructed based on a unified association identifier. Synchronization verification records can be stored in the form of structured data tables, key-value mapping structures, or multi-dimensional state matrices.

[0030] A compliant recommended device refers to a fully compliant device entity that meets the preset matching status across all constraint dimensions and whose specific parameters reach the preset performance threshold. Here, "all constraint dimensions" emphasizes the simultaneous fulfillment of all rigid conditions by the same device entity under the same task, rather than the concatenation result of multiple devices meeting different conditions separately. The preset performance threshold emphasizes that, in addition to the matching relationship, the specific performance indicators carried in the relationship edges or node attributes need to be verified at a minimum. For example, even if a device has a connectivity relationship with the target power node, if its actual usable capacity is lower than the project requirements, it still cannot be considered a compliant recommended device.

[0031] In practice, obtaining multiple constraint dimensions of requirements for electrical drive equipment can be achieved by parsing the project requirement document entered by the sales engineer. For example, the project requirement document can come from a form entry page in the sales management system or from an uploaded electronic document. If it comes from a form entry page, fields such as power requirements, protection requirements, and certification requirements can be directly read; if it comes from an electronic document, the document can first be extracted, and then segmented and categorized according to preset requirement keyword rules.

[0032] More specifically, when retrieving candidate targets based on a pre-defined relationship model, a relational database table can be constructed. A database model is pre-built, including a basic equipment information table, a power parameter table, a protection level table, and a certification qualification table. The basic equipment information table stores basic fields such as equipment number, equipment model, product series, and applicable industry; the power parameter table stores the correspondence between equipment number and fields such as rated power parameters, overload capacity description, and compatible load type; the protection level table stores the correspondence between equipment number and fields such as IP protection level, environmental compatibility description, and anti-corrosion treatment description; and the certification qualification table stores the correspondence between equipment number and fields such as certification name, certification status, certificate number, and validity period. This database model can be formed by organizing the company's existing product master data, technical parameter database, and qualification management ledger. During organization, the equipment number can be used as a unified primary key to clean and align parameter information originally scattered across different business systems before importing it into a unified data storage structure. Query statements are then executed in each of these tables.

[0033] Furthermore, the pre-defined relationship model can also include field standardization and relationship mapping establishment. For example, a batch of records showing the correspondence between equipment models and parameter fields can be manually compiled first, and then relationships such as equipment number—power parameter, equipment number—protection level, and equipment number—certification qualification can be written into the database to obtain a relationship model that can be retrieved and called.

[0034] In some implementations, generating synchronization verification records based on association identifiers can be achieved by establishing a data mapping dictionary. The device ID of the candidate target is used as the key, and a list containing multiple strings is used as the value. When a local matching state with an identifier is identified, it is appended to the list of corresponding device IDs as a string.

[0035] To make the process more stable and feasible, a temporary cache structure with a combined index of the association identifier and device number can be established first during the actual writing process. Then, the data is written to the corresponding positions according to the predetermined order required by the constraint dimensions. Therefore, in a more specific implementation, position mappings can be generated first based on the dimension definitions in the task. For example, the power dimension can be mapped to the first position, the environment dimension to the second position, and the authentication dimension to the third position. Subsequently, when a local matching status is returned, its corresponding dimension is identified first, and then the data is written to the corresponding position. If no matching result is found for a certain dimension, the position is left empty or a preset miss flag is written. The synchronous verification records generated in this way not only reflect the status of the device in each dimension but also ensure that the record structure between different devices is completely consistent, facilitating subsequent unified comparison.

[0036] Finally, these synchronous verification records are traversed, and each indicator in the record is compared with preset performance thresholds (e.g., minimum power 500kW) and matching status (e.g., certification is required). Only when all dimensions in the record meet the requirements is the candidate target identified as a compliant recommendation. For the power dimension, the system checks whether there is a valid local power matching status in the synchronous verification records and determines whether the performance indicator attribute value corresponding to that status meets the project requirements. For the environment dimension, the system checks whether there is a protection level or environment adaptation status that meets the requirements. For the certification dimension, the system checks whether there is a valid certification hit status. If any dimension is missing, misses a match, or has a record but does not meet the preset requirements, the candidate target will not be identified as a compliant recommendation.

[0037] Preset performance thresholds and matching states can be established through rule configuration before deployment. For example, the sales management platform can provide a rule maintenance interface, allowing technical administrators to input judgment rules corresponding to common project requirements into the rule base, forming rules such as numerical comparison for power-related requirements, existence verification for certification-related requirements, and level matching or adaptation specification verification for environmental-related requirements. When the system performs the final judgment, it does not require temporary manual interpretation of each requirement, but directly calls the pre-configured rules for unified processing. In this way, the final output of compliant recommended objects not only meets the multi-dimensional joint requirements, but also its judgment process is traceable and consistent.

[0038] The above technical solution establishes multi-dimensional constraints belonging to the same business entity from the data input stage. During the candidate retrieval phase, the independent matching status of each dimension is objectively preserved, and the scattered local matching information is aggregated into a synchronous verification record for the device entity through a unified association identifier. In the final recommendation decision, joint verification and threshold comparison covering all dimensions are performed based on this record. This completely solves the technical defect of traditional isolated reasoning that may mistakenly recommend non-compliant devices due to single-condition hits, ensuring that the output compliant recommendations can simultaneously meet complex operating conditions and industry access standards. This significantly improves the technical feasibility and accuracy of the sales selection results.

[0039] In some preferred embodiments, the internal topological structure of the association model and the traversal retrieval process based on the model are further refined. The specific technical solutions are as follows: The relationship model is a knowledge graph, whose nodes include equipment model nodes and performance nodes. The performance nodes include power segment nodes, environmental protection level nodes, corrosive environment adaptability nodes, and industry certification nodes. The nodes are connected by static relationship edges, which carry performance indicator attribute values, including rated power values, overload capacity percentage, and IP protection level parameters. Based on a pre-defined association model, candidate targets that match the requirements of each constraint dimension are retrieved, and the local matching status between each candidate target and the requirements of each constraint dimension is recorded, including: Starting from the performance nodes corresponding to the requirements of each constraint dimension, we traverse and reason along the static relationship edges in the relationship model to the device model nodes to determine the candidate objects. Record the static relation edges that the association model passes through during the process of determining candidate objects, and use the performance index attribute values ​​in the static relation edges as local matching states.

[0040] In this specific implementation plan, the knowledge graph constructs a complete entity topology network, abstracting power segments, environmental protection levels, corrosive environment adaptability, and industry certifications as independent performance nodes, with device models serving as target entity nodes. These nodes are connected via static relationship edges. These static relationship edges not only establish topological associations but also carry refined parameters such as rated power values, overload capacity percentages, and IP protection levels in their attribute structures. For certification-related relationship edges, they can further carry certification names, certification status, and validity identifiers; for corrosive environment adaptability relationship edges, they can also carry descriptions of the adaptability environment category or anti-corrosion treatment methods.

[0041] To ensure the feasibility of these pre-defined graph relationships, a combination of manual verification and system import is preferred: first, equipment models and performance parameters are extracted from equipment technical data; then, the data maintenance module imports them into the graph database according to a unified field template; finally, sampling verification confirms the accuracy of the relationships between equipment nodes and performance nodes. During retrieval, the performance nodes corresponding to the requirements of each constraint dimension are used as the seed starting point for graph traversal. Graph traversal and diffusion reasoning are performed backward along the static relationship edges towards the equipment model nodes. All reachable equipment model nodes are captured as candidate objects. Simultaneously, during the traversal, the specific performance index attribute values ​​on the traversed static relationship edges are captured and recorded in real time, and directly converted into local matching states. Reverse traversal along the static relationship edges does not require the edge directions in the graph database to be fixed in a certain form. As long as the reachable path rules from performance nodes to equipment model nodes are clearly defined in the retrieval configuration, the effect of recalling equipment starting from constraints can be achieved. In this way, while recalling candidate devices, the attribute basis required for subsequent joint discrimination is obtained simultaneously, avoiding the additional data jumps caused by recalling first and then querying attributes a second time.

[0042] In a specific example, for the heavy-load operating conditions and protection levels required in the project, the heavy-load power node and protection level node in the knowledge graph are set as the traversal starting point. The graph traversal engine spreads along the static relationship edge of the adaptation power to the device model node, hitting the device model node G, and simultaneously reading the attribute values ​​of rated power and overload capacity on this relationship edge; at the same time, another traversal path spreads along the shell protection relationship edge, also hitting G, and reading the attribute value of protection level on the edge. These specific attribute values ​​extracted from the static edges are recorded as the local matching state of device G in the power dimension and protection dimension. If the project also includes corrosive environment adaptation and industry certification requirements, the same traversal continues from the corresponding performance node, and the hit results are written into the other dimension states of the same device; if a dimension is not hit, the missing state is explicitly recorded in that dimension, instead of being ignored by default. In this way, when constructing the subsequent synchronous verification records, device G already has a complete multi-dimensional state foundation.

[0043] In some preferred embodiments, the aggregation and structured construction mechanism of multidimensional constraint local matching states is further refined, and the specific technical solutions are as follows: The steps for generating synchronization verification records by assigning corresponding local matching states to each candidate target based on the association identifier include: Based on the association identifier, a multi-dimensional state matrix is ​​created for each candidate target, and the dimensions of the multi-dimensional state matrix correspond one-to-one with the multiple constraint dimensions requirements; The retrieved local matching states are filled into the corresponding dimensions of the multidimensional state matrix to form a synchronous verification record of the candidate target for the electrical drive equipment.

[0044] In this specific implementation scheme, based on a globally unique association identifier, a dedicated multidimensional state matrix is ​​instantiated in computational memory for each recalled candidate target. The axis or column indices of each dimension of the multidimensional state matrix strictly adhere to the dimensional requirements of the project's multidimensional constraint set. A dimension template can be generated first based on the current task's constraint dimension list, and then a corresponding record can be created for each candidate device based on this template. Subsequently, a structured mapping filling operation is performed, accurately writing the various local matching states (including qualitative hit states and quantitative performance attribute values) recorded during the graph traversal into the corresponding dimension positions in the multidimensional state matrix. During filling, it is preferable to simultaneously write the following: dimension name, hit state, attribute value summary, relationship source identifier, and association identifier.

[0045] For candidate devices that are not traversed and hit on certain dimensional paths, write null values ​​or non-compliance flags in the corresponding dimensions of the matrix. If there are multiple candidate hit relationships in a certain dimension, such as multiple authentication records for the same device, valid authentication records can be retained according to preset priority rules, or multiple hit results can be written side by side into the extended field of that dimension, and then filtered by the subsequent joint discrimination module according to the validity rules.

[0046] In some preferred embodiments, after determining compliant recommendation targets from candidate targets, the method further includes: For candidate targets that have not been identified as compliant recommendation objects, retrieve dimension items with matching status of "not met" or "not matched" from the synchronous verification records; The dimension item is identified as the missing constraint dimension requirement, and based on the preset business rule table and preset substitutability rules, it is determined whether the missing constraint dimension requirement belongs to a substitutable constraint that can be made up by external means. If the missing constraint dimension requirement is a replaceable constraint, then a compensation scheme suggestion is generated for the candidate object; If the missing constraint dimension requirement is not a replaceable constraint, then a permanent block is applied to the candidate object.

[0047] In this specific implementation plan, the handling of non-compliant candidate devices is based on a deep reverse lookup of the synchronous verification records. By traversing the multi-dimensional state matrix of the non-selected candidate devices, the specific dimension item that caused the joint judgment failure is located and marked as the missing constraint requirement of the current device in this project task. Here, the traversal optimization is to check each device and each dimension according to the project combination affiliation identifier. For example, the power segment matching dimension, environmental protection dimension, corrosive environment adaptability dimension, and industry certification dimension are checked in sequence, and the corresponding status field, original attribute value field, and hit path field of each dimension are read simultaneously. If the status of a certain dimension is a miss, the source of the missing reason for that dimension is further read, such as whether there is no corresponding relationship edge in the graph, the device attribute value is lower than the project threshold, the certification validity period has expired, or the device model is not configured with the corresponding structural component at all.

[0048] Subsequently, the pre-configured business rule table and alternative rule base are invoked to perform classification and determination. The business rule table defines in detail the physical and engineering attribute classifications of each constraint dimension, while the alternative rules pre-determine whether external compensation measures are allowed for each type of constraint and their access boundaries. The business rule table can be compiled from the company's historical project experience, product design boundaries, certification compliance requirements, and after-sales maintenance feedback. Alternative rules can be established by manually identifying the correspondence between typical missing items and feasible compensation measures. For example, mapping insufficient environmental protection level to the addition of external high-protection-level cabinets; mapping insufficient adaptation to mildly corrosive environments to the addition of anti-corrosion covers or anti-corrosion sealed electrical control cabinets; and mapping the lack of access permission for the target standard to non-alternative measures. Each rule in the rule base also preferably includes applicable prerequisites, such as the applicable environmental level range, the types of attachments allowed for compensation, whether secondary verification is required, and whether it is allowed to enter the manual review queue. For example, missing anti-corrosion coatings or insufficient environmental protection levels are typically defined as constraints with engineering substitutability, while industry access qualifications and topology limitations are defined as inherent, irreplaceable rigid constraints of the equipment itself. When a missing item is determined to be a substitutable constraint, the system automatically generates corresponding compensation suggestions (such as adding a specified model of enclosure protection component or external filter unit) and marks the candidate equipment as compensable.

[0049] The compensation plan should include at least the name of the compensation attachment, the scope of attachment specifications, applicable operating conditions, installation location, and a prompt indicating that secondary verification is required. If multiple optional compensation attachments exist in the rule base, the priority should be selected based on the principle of best matching the current project environment, having a high historical success rate, or having a short delivery cycle. Once an irreplaceable rigid constraint is detected as missing, a blocking mechanism should be triggered to permanently terminate the recommended circulation of that equipment in the current project. This blocking action should preferably be simultaneously recorded in the blocking record table, with the record including at least the equipment identifier, project portfolio affiliation identifier, blocking dimension, blocking reason text, trigger time, and rule source number, for traceability during subsequent audits, manual review, and rule revisions.

[0050] In some preferred embodiments, after generating compensation scheme suggestions for candidate objects, the method further includes: Obtain the secondary impact correction coefficient corresponding to the compensation plan suggestion. The secondary impact correction coefficient is used to characterize the negative physical impact of the compensation accessories on the performance of the equipment itself. The original performance index of the candidate target is weighted and calculated based on the secondary impact correction coefficient to obtain the corrected performance index. The corrected performance index is equal to the product of the original performance index and the secondary impact correction coefficient. The revised performance metrics are compared with the performance thresholds a second time. If the corrected performance metrics exceed the performance threshold, the original compensation plan should be maintained. If the corrected performance metrics fall below the performance threshold, the proposed compensation scheme will be revoked, and the corresponding candidate targets will be permanently blocked.

[0051] In this specific implementation plan, a secondary impact mapping table of compensation accessories can be pre-built and maintained. This table records in detail the quantitative reduction ratio of various key performance indicators of the equipment body by various external compensation accessories under different operating conditions, i.e., the secondary impact correction coefficient.

[0052] This mapping table is feasible. Preferably, the data sources in the mapping table include the following categories: first, test records generated after thermal, voltage, and load tests on typical accessory and equipment combinations during the R&D phase; second, operational monitoring data after actual accessory installation in historical projects; third, derating recommendations explicitly given in product manuals or certification documents; and fourth, experience-based correction values ​​confirmed by engineers. The preferred record format of the mapping table includes fields such as accessory type, accessory model or specification range, applicable ambient temperature range, applicable installation method, name of performance-affecting index, secondary impact correction coefficient, data source, and activation status. For example, a heat dissipation derating coefficient (e.g., 0.85) is defined for a sealed dustproof cabinet, and a voltage attenuation coefficient (e.g., 0.97) is defined for an incoming line reactor. When a compensation scheme suggestion is generated, the candidate device is automatically suspended, the specific accessory types involved in the compensation scheme are extracted, and the corresponding secondary impact correction coefficient is retrieved from the mapping table. Suspension here means temporarily placing the device in a state awaiting secondary verification, preventing it from directly entering the recommended queue.

[0053] During the retrieval process, it is preferable to consider contextual information such as attachment type, attachment specifications, ambient temperature, installation location, and ventilation conditions to avoid incorrectly applying inapplicable correction coefficients. Subsequently, the original performance indicators (e.g., rated continuous output power, maximum overload current, etc.) registered in the synchronous verification record of the equipment are retrieved, and the corrected performance indicators after adding attachments are calculated through multiplicative weighted summation. The original performance indicators here are preferably directly derived from the equipment attribute values ​​extracted and verified in the preceding joint verification record to ensure data consistency. If a compensation scheme includes multiple attachments, they can be applied sequentially to the corresponding performance indicators according to the preset combination rules in the mapping table, and the results of each step are written back to the temporary verification record to track the impact of each attachment.

[0054] Next, the comparison engine will re-compare the corrected performance metrics with the performance thresholds in the joint constraint set for this non-standard project. During the comparison, the same threshold criteria as the previous rigid constraint verification are preferred. For example, power metrics are compared with the project's power threshold, voltage metrics with the project's minimum allowable voltage, and overload capacity with the project's heavy-load requirements. If the corrected metrics still exceed the thresholds, it proves that the equipment's performance margin is sufficient to absorb the secondary losses caused by the accessories, confirming the feasibility of the compensation scheme and maintaining it. If the corrected metrics fall below the thresholds, it indicates that the equipment cannot meet the core operating load after additional compensation, and the compensation suggestion is immediately revoked and a permanent shutdown is enforced. During revocation, the equipment status is preferably updated simultaneously, rewriting the original compensable flag as indicating non-compliance due to secondary impacts of compensation, and generating new shutdown reason text so that downstream systems and sales interfaces can accurately display the information.

[0055] In one specific embodiment, for the aforementioned candidate device M, its original rated power is 45kW, and the minimum rated power threshold required by the current project conditions is 37kW. The compensation suggestion generated for this device is to install an external IP55 anti-corrosion sealed electrical control cabinet. The secondary impact mapping table of the compensation appendix is ​​retrieved, and the heat dissipation derating factor for this sealed cabinet at an ambient temperature of 45 degrees Celsius is found to be 0.80. To ensure the correct use of this factor, it is preferable to also verify whether the installation method of the sealed cabinet is consistent with the applicable conditions in the mapping table, such as whether it is a wall-mounted sealed cabinet, whether natural heat dissipation is used, or whether it is in a high-temperature and high-humidity environment. Only when the applicable conditions are consistent is the correction factor used for weighted calculation: corrected rated power = 45kW × 0.80 = 36kW. After the calculation is completed, the corrected power value of 36kW is compared a second time with the project's minimum threshold of 37kW, identifying that the corrected power is lower than the project's minimum requirement (36kW < 37kW). This indicates that although device M compensates for its protection level deficiencies with an external enclosed enclosure, the resulting temperature rise and thermal derating prevent its actual usable power from driving the load. A secondary blocking branch is immediately triggered, revoking the enclosure compensation scheme recommendation for M, marking it as a power non-compliance due to secondary derating, and implementing permanent blocking. Conversely, if the device's original rated power is 55kW and its corrected power is 44kW (55kW × 0.80 = 44kW > 37kW), then its surplus power is deemed sufficient to cover heat dissipation losses, and the compensation scheme recommendation is officially retained. In this example, the secondary verification results can also be written into the compensation scheme closed-loop record table, recording the original power, accessory type, applicable ambient temperature, correction coefficient used, corrected power, comparison threshold, and final processing conclusion for subsequent project review and rule optimization.

[0056] In some preferred embodiments, after determining compliant recommendation targets from candidate targets, the method further includes: Based on the judgment results, priority tiers are divided, with compliant recommended objects classified as the first priority tier and candidate targets with alternative constraints classified as the second priority tier. Extract the performance metrics of the compliant recommended objects under the requirements of various constraint dimensions, and calculate the redundancy coefficient of the performance metrics relative to the preset bottom line requirements; Based on the redundancy coefficient and the preset historical sales conversion rate, calculate the comprehensive recommendation score of the compliant recommended objects, and sort the objects in the first priority tier in descending order according to the comprehensive recommendation score; Perform cost penalty deduction on candidate targets in the second priority tier and sort them in descending order after the ranking results of the first priority tier; Based on the order of the first priority tier and the second priority tier, generate and output the recommended ranking results.

[0057] In this specific implementation plan, firstly, feasible devices are strictly isolated into physical tiers based on their compliance status: fully compliant recommended devices enter the first priority tier, while compliant devices requiring additional external compensation measures enter the second priority tier. This tier division is preferably based directly on prior status indicators, i.e., checking whether the device has a joint establishment marker, whether it has a compensation plan suggestion, whether it has passed secondary verification, and whether it has a permanent blocking record. Only devices that simultaneously meet the conditions of joint establishment and do not require compensation enter the first priority tier; only devices that meet the conditions of having alternative constraints, having generated a compensation plan suggestion, and having passed secondary verification enter the second priority tier.

[0058] Within the first priority tier, the technological advancement and adaptability of equipment are reflected by quantifying its performance margin across various constraint dimensions. Specifically, the difference ratio between the actual parameters of the equipment and the project's minimum requirements is extracted, and indicators such as power redundancy and protection redundancy are calculated and combined into a redundancy coefficient. Simultaneously, the sales conversion rate of this equipment model in similar historical industry projects is introduced. The technological margin and market acceptance are weighted and summed to obtain a comprehensive recommendation score for the first-tier equipment, which is then ranked in descending order of score. The historical sales conversion rate is preferably derived from historical project data in the enterprise's sales management system. The statistical scope can be limited to transactions within the same industry, similar power range, similar environmental level, or the same regional market to avoid data interference between different business scenarios. To ensure the feasibility of the scoring process, a scoring parameter configuration table can be maintained in advance, recording the weight of redundancy in each dimension, the weight of historical sales conversion rate, and whether certain scoring items are enabled under different project types.

[0059] For equipment in the second priority tier, since they require the procurement and installation of additional compensation accessories (such as protective cabinets and filters), which increases the project's material costs and delivery cycle, a cost penalty is enforced based on their basic performance score. The second-tier equipment, after the deductions, is then ranked after all equipment in the first priority tier, ultimately generating a globally ordered recommended equipment sequence. This cost penalty deduction comprehensively considers factors such as accessory procurement costs, installation time, extended delivery cycles, and increased maintenance complexity. Enterprises can preset different deduction levels according to their business strategies; for example, lower deductions for low-cost accessories and higher deductions for high-cost and complex-to-install accessories. This approach not only reflects technical feasibility but also the commercial implementation costs.

[0060] In one specific embodiment, during a product selection task, both Equipment A and Equipment B are fully compliant and recommended, and are placed in the first priority tier. Equipment C, a device with a sealed cabinet compensation that has passed secondary verification, is placed in the second priority tier. For the first tier, the project's required power is 55kW. Equipment A has a rated power of 75kW, a power redundancy coefficient of (75-55) / 55=0.364, and a historical sales conversion rate of 0.85. Equipment B has a rated power of 60kW, a power redundancy coefficient of (60-55) / 55=0.091, and a historical sales conversion rate of 0.90. Based on preset weights (e.g., redundancy weight 0.6, conversion rate weight 0.4), A's overall recommendation score is higher than B's. Therefore, A ranks first and B second in the first tier.

[0061] The scoring process preferably begins by retrieving the rated power, protection level, and certification status of both devices from the joint verification record to confirm that they meet the project's minimum requirements. Next, it retrieves transaction records for both devices in similar projects from the historical sales database to calculate the historical sales conversion rate. Finally, it uses the weighted parameters from the scoring configuration table to complete the comprehensive scoring. For device C in the second tier, after calculating its basic technical score, a 20-point penalty for materials and labor costs due to the installation of a sealed cabinet is deducted to complete the internal ranking within the second tier. This penalty score is preferably derived from a preset compensation cost rule table, which can record, for example, sealed cabinet accessories, cost penalty levels, installation complexity levels, and default deduction values. The final recommended sequence output by the system strictly reflects: Tier 1, first-tier A; Tier 1, second-tier B; Tier 2, C (with compensation suggestions). During output, the system also preferably displays each device's tier affiliation, comprehensive recommended score, main advantages, and whether compensation suggestions are included, so that sales personnel can quickly understand the reasons for the ranking.

[0062] In some preferred embodiments, the step of calculating the comprehensive recommendation score of compliant recommended objects based on the redundancy coefficient and the preset historical sales conversion rate includes: Obtain historical maintenance records and component failure rates of compliant recommended objects, and combine them with preset equipment deployment environment parameters to determine the performance degradation prediction curve of compliant recommended objects; Based on the performance degradation prediction curve, calculate the performance envelope area of ​​the compliant recommended object within the preset service life. The redundancy coefficient is corrected based on the area of ​​the performance envelope to obtain the time redundancy score, which is the product of the area of ​​the performance envelope and the redundancy coefficient. The timeliness redundancy score is weighted with the historical sales conversion rate to obtain a comprehensive recommendation score for compliant recommended targets.

[0063] In this specific implementation plan, the historical maintenance records of the target equipment model in the enterprise quality database (such as mean time between failures (MTBF) and mean time to repair (MTTR)) and the standard failure rate data of each key power component (such as thermal failure rate parameters under the Arrhenius model) are retrieved first. Combined with the actual deployment environment parameters input by the current non-standard project (such as annual average ambient temperature, relative humidity, power grid harmonic distortion rate, etc.), a dynamic function characterizing the availability performance of the equipment over time during its future service life is constructed, namely the performance degradation prediction curve.

[0064] Specifically, historical maintenance records are derived from the company's after-sales service system, repair record system, and quality tracking system; component failure rates are obtained from component supplier reliability manuals, internal life test reports, or standard reliability databases; and equipment deployment environment parameters are derived from project requirements entered by sales engineers, on-site survey reports, or operating condition specifications provided by customers. When constructing performance degradation prediction curves, key components affecting the lifespan of the equipment model are identified first, and then corresponding degradation templates are selected based on factors such as ambient temperature, humidity, corrosion level, and power grid quality. For example, for equipment operating for extended periods in high-temperature environments, the degradation weight of components related to heat dissipation can be increased; for high-humidity and high-corrosion environments, the degradation weight related to insulation and connector aging can be increased. Subsequently, a definite integral is calculated on the performance degradation prediction curve over a preset service life (e.g., a standard service life of 8 years) to obtain the area of ​​the performance envelope below the curve. This step is preferably implemented using discrete-time sampling, i.e., reading the performance values ​​of the prediction curve at each time point by year, quarter, or month, and then accumulating them to obtain the total equivalent performance over the entire service life, thus avoiding reliance on complex continuous calculations at the implementation level. This envelope area quantitatively represents the total equivalent effective performance that the equipment can continuously output throughout its entire life cycle; a larger area indicates stronger anti-aging capabilities and better long-term stability. This envelope area is used as a timeliness correction factor and multiplied by the initial static redundancy coefficient to calculate the timeliness redundancy score. Finally, the timeliness redundancy score is weighted and averaged with historical sales conversion rates to obtain the final comprehensive recommendation score. The weighting here is preferably configured by the enterprise according to business preferences; for example, increasing the weight of the timeliness redundancy score in industry projects emphasizing long-term reliability, and appropriately increasing the weight of historical sales conversion rates in regular projects emphasizing rapid transactions.

[0065] In one specific embodiment, both compliant equipment H and equipment S have an initial power redundancy coefficient of 0.30 at the time of manufacture. The project's preset service life is 10 years, and the site environment is an average annual high temperature of 40 degrees Celsius. The failure rate and historical maintenance data of the core components of the two devices are obtained: Equipment H uses industrial-grade long-life film capacitors and a high-specification heat dissipation topology, and its performance degradation prediction curve under high-temperature conditions is [curve description needed]; Equipment S uses ordinary electrolytic capacitors, and its performance degradation prediction curve is [curve description needed]. In this example, the preferred approach is not to directly judge which is more durable based on experience, but rather to first read the repair frequency, fan replacement cycle, capacitor failure records, and on-site downtime records of both models in the high-temperature project from the quality database, and then combine this with the differences in component selection to establish their respective degradation prediction curves. The system performs integral calculations on the envelope areas of the two curves within the 0 to 10-year interval, and after normalization, the performance envelope area of ​​equipment H is 0.93, while the performance envelope area of ​​equipment S is 0.78. Calculate the time redundancy score separately: Time redundancy score of equipment H = 0.30 × 0.93 = 0.279; Time redundancy score of equipment S = 0.30 × 0.78 = 0.234.

[0066] The time-related redundancy score was then weighted and fused with the respective historical conversion rates. Because device H has stronger anti-aging capabilities and a larger full-cycle performance envelope area, its final comprehensive recommendation score was significantly higher than that of device S, and it received priority recommendation in the first priority tier.

[0067] In some preferred embodiments, the method further includes: When none of the candidate targets are identified as compliant recommendations, all candidate targets subject to permanent blocking will be diverted to the manual review queue. Cluster analysis is performed on the reasons for the missing candidate targets in the manual review queue to generate an early warning report; Obtain feedback from the manual review queue, and update the entity relationships in the association model based on the feedback, or adjust the parsing rule parameters of the constraint dimension requirements.

[0068] In this specific implementation plan, when the joint judgment module and the secondary verification module have scanned all candidate devices under the current task and confirmed that no device can be identified as a compliant recommendation object, the process is not terminated directly. Instead, the exception fallback logic is triggered, and all candidate targets corresponding to this task, their associated synchronous verification records, and detailed reason tags for triggering permanent blocking are packaged together and automatically distributed to the manual review queue of the sales management system.

[0069] Subsequently, the backend analysis engine extracts the missing cause data for all blocking devices in the manual review queue and performs attribution cluster analysis. This cluster analysis can employ high-frequency itemset mining based on feature frequency statistics, or algorithms such as K-Means and text semantic similarity clustering, to summarize the bottleneck constraints leading to overall blocking into typical contradiction categories that erupt in concentrated bursts (e.g., widespread lack of tolerance to ultra-low temperature environments or a 100% missing explosion-proof certificate rate in a specific industry), and automatically generates a structured selection blocking early warning report.

[0070] Furthermore, after the warning report is pushed to system maintenance personnel, senior application engineers, or product line experts, the system monitors the interactive feedback on the manual review interface and feeds back the processing conclusions confirmed by the human reviewers to the underlying algorithm engine. Specifically: if the manual review feedback indicates that the blocking reason is missing data in the knowledge graph (for example, a series of devices has recently completed a technical upgrade and obtained target standard access certification, but the relevant static relationship edges have not yet been entered into the database), then the entity relationship edges between the device model node and the corresponding performance node are dynamically added to the relationship model (knowledge graph) according to the human review instructions; if the manual review feedback indicates that the blocking reason is that the rules set in the requirement parsing stage are too stringent (for example, the vague description proposed by the customer is misjudged by the natural language parsing module as an extremely strict rigid indicator), then the parsing rule threshold or keyword mapping weight for the requirement of that constraint dimension is adjusted accordingly.

[0071] In some preferred embodiments, before determining compliant recommendation objects from candidate targets based on synchronization verification records, preset matching states, and preset performance thresholds, the method further includes: Identify dimension pairs that are physically or logically coupled among multiple constraint dimensions and use them as a collaborative verification dimension group; The interaction interference coefficients of candidate targets under the collaborative verification dimension group are extracted from the correlation model. The interaction interference coefficients are used to characterize the performance loss when different dimensional requirements are implemented simultaneously on the same hardware carrier. Based on the cross-interference coefficient, the corresponding local matching state in the synchronous verification record is reduced to generate a collaborative correction record; Based on synchronous verification records, preset matching status, and preset performance thresholds, compliant recommended targets are determined from candidate targets, including: Candidate targets are jointly determined based on synchronous verification records and collaborative correction records. A candidate target is determined to be a compliant recommendation object if and only if the corrected state of the candidate target after considering the interference between dimensions is still in a matching state and meets the performance threshold.

[0072] In this specific implementation scheme, a multi-dimensional physical interference collaborative verification logic is added before the final judgment is performed. First, the multi-dimensional constraint list of this task is retrieved. Based on the preset physical coupling rule base, the combination of dimensions that have physical field constraints or logical relationships with each other is scanned and identified, and these are packaged and defined as collaborative verification dimension groups (e.g., [heavy load power demand, high grid carrier frequency demand], [high altitude low air pressure environment, continuous rated current demand], etc.).

[0073] Subsequently, specific attributes of the target device model under the corresponding collaborative verification dimension group are retrieved from the association model (knowledge graph), and the cross-interference coefficients, which have been pre-determined through type testing or thermal simulation calculations, are extracted. These cross-interference coefficients precisely quantify the mutual attenuation ratio of core parameters when two or more dimensions are simultaneously under high load.

[0074] Next, the calculation engine calls the interactive interference coefficient to perform multiplication or derating on the static local matching state parameters (such as rated power, rated current, maximum allowable ambient temperature, etc.) originally recorded in the synchronous verification record, generating a collaborative correction record that reflects the true limit capability under the coexistence state of multidimensional physical fields.

[0075] Finally, the joint judgment module will conduct a dual linkage judgment on the synchronous verification record and the collaborative correction record: it not only requires that the candidate device's original local matching state in each independent dimension reaches the baseline, but also requires that its collaborative correction state after being subjected to interactive interference reduction can still fully cover the preset performance threshold bottom line; only when both judgments pass, the candidate target is finally given the identity of a compliant recommendation object.

[0076] Secondly, referring to Figure 2 This application also proposes a knowledge graph-based sales management system for electrical drive equipment, used to perform the steps in the aforementioned method, including: The requirement acquisition module 210 is used to acquire multiple constraint dimension requirements of electrical drive equipment and configure a unified association identifier for each constraint dimension requirement; The retrieval record module 220 is used to retrieve candidate targets that match the requirements of each constraint dimension based on a preset association relationship model, and record the local matching status between each candidate target and the requirements of each constraint dimension. The record building module 230 is used to assign corresponding local matching states to each candidate target according to the association identifier in order to generate a synchronization verification record; The joint judgment module 240 is used to determine compliant recommended objects from candidate targets based on synchronous verification records, preset matching status, and preset performance thresholds.

[0077] Through the above technical solution, the functional modules work collaboratively. The requirement acquisition module 210 establishes task-level association bindings with full-dimensional constraints at the data entry point; the record retrieval module 220 uses a knowledge graph to perform efficient topological traversal and preserve local matching traces; the record construction module 230 aligns and assembles multi-source discrete states into standardized synchronous verification records; and the joint judgment module 240 implements veto and threshold bottom-line verification based on this global verification entity. The system completely solves the defects of incorrect and missed reasoning caused by discrete single-path reasoning from the perspective of data structure and processing flow, ensuring that the output device meets the full-dimensional strong constraints of complex engineering conditions and qualification restrictions.

[0078] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A sales management method for electrical drive equipment based on knowledge graphs, characterized in that, The steps of this method include: Obtain multiple constraint dimension requirements of the electrical drive equipment, and configure a unified association identifier for each constraint dimension requirement; Based on a pre-defined association model, candidate targets that match the requirements of each constraint dimension are retrieved, and the local matching status between each candidate target and each constraint dimension requirement is recorded. Based on the association identifier, each candidate target is assigned a corresponding local matching state to generate a synchronization verification record; Based on the synchronous verification records, the preset matching status, and the preset performance threshold, compliant recommended objects are determined from the candidate targets.

2. The sales management method for electrical drive equipment based on knowledge graphs according to claim 1, characterized in that, The association model is a knowledge graph, whose nodes include equipment model nodes and performance nodes. The performance nodes include power segment nodes, environmental protection level nodes, corrosive environment adaptability nodes, and industry certification nodes. The nodes are connected by static relationship edges, which carry performance indicator attribute values, including rated power values, overload capacity percentage, and IP protection level parameters. The method, based on a preset association model, retrieves candidate targets that match the requirements of each constraint dimension, and records the local matching status between each candidate target and each constraint dimension requirement, including: Starting from the performance node corresponding to each of the aforementioned constraint dimensions, traverse and reason along the static relationship edge in the association model to the device model node to determine candidate objects; Record the static relationship edges that the association model passes through in the process of determining the candidate object, and use the performance index attribute values ​​in the static relationship edges as the local matching state.

3. The sales management method for electrical drive equipment based on knowledge graphs according to claim 1, characterized in that, The step of assigning corresponding local matching states to each candidate target based on the association identifier to generate a synchronization verification record includes: Based on the association identifier, a multidimensional state matrix is ​​created for each candidate target, and the dimensions of the multidimensional state matrix correspond one-to-one with the multiple constraint dimension requirements; The retrieved local matching states are filled into the corresponding dimensions of the multidimensional state matrix to form a synchronous verification record of the candidate target for the electrical drive equipment.

4. The sales management method for electrical drive equipment based on knowledge graphs according to claim 1, characterized in that, After determining the compliant recommendation targets from the candidate targets, the method further includes: For candidate targets that are not identified as compliant recommendation objects, retrieve dimension items in the synchronous verification record whose matching status is not met or not matched. The dimension item is identified as a missing constraint dimension requirement, and based on the preset business rule table and preset substitutability rules, it is determined whether the missing constraint dimension requirement belongs to a substitutable constraint that can be compensated by external means. If the missing constraint dimension requirement belongs to the replaceable constraint, then a compensation scheme suggestion is generated for the candidate object; If the missing constraint dimension requirement is not among the alternative constraints, then the candidate object is permanently blocked.

5. The sales management method for electrical drive equipment based on knowledge graphs according to claim 4, characterized in that, After generating compensation scheme suggestions for the candidate objects, the method further includes: Obtain the secondary impact correction coefficient corresponding to the proposed compensation scheme. The secondary impact correction coefficient is used to characterize the negative physical impact of the compensation accessories on the performance of the equipment itself. The original performance index of the candidate target is weighted and calculated according to the secondary impact correction coefficient to obtain the corrected performance index, which is equal to the product of the original performance index and the secondary impact correction coefficient. The corrected performance index is compared with the performance threshold a second time; If the corrected performance index is higher than the performance threshold, then the original compensation scheme recommendation shall be maintained. If the corrected performance metric is lower than the performance threshold, the compensation scheme recommendation is revoked, and the corresponding candidate target is permanently blocked.

6. The sales management method for electrical drive equipment based on knowledge graphs according to claim 4, characterized in that, After determining the compliant recommendation targets from the candidate targets, the method further includes: Based on the judgment results, a priority tier division is performed, classifying the compliant recommended objects into the first priority tier and classifying the candidate targets with substitutable constraints into the second priority tier. Extract the performance indicators of the compliant recommended objects under each of the constraint dimensions, and calculate the redundancy coefficient of the performance indicators relative to the preset bottom line requirements; Based on the redundancy coefficient and the preset historical sales conversion rate, calculate the comprehensive recommendation score of the compliant recommended objects, and sort the objects in the first priority tier in descending order according to the comprehensive recommendation score; The candidate targets in the second priority tier are subject to cost penalty deduction and are then sorted in descending order after the ranking results of the first priority tier. Based on the order of the first priority tier and the second priority tier, generate and output the recommended ranking result.

7. The sales management method for electrical drive equipment based on knowledge graphs according to claim 6, characterized in that, The step of calculating the comprehensive recommendation score of the compliant recommended object based on the redundancy coefficient and the preset historical sales conversion rate includes: Obtain the historical maintenance records and component failure rates of the compliant recommended objects, and combine them with preset equipment deployment environment parameters to determine the performance degradation prediction curve of the compliant recommended objects; Based on the performance degradation prediction curve, calculate the performance envelope area of ​​the compliant recommended object within the preset service life. The redundancy coefficient is corrected based on the performance envelope area to obtain a time redundancy score, which is the product of the performance envelope area and the redundancy coefficient. The timeliness redundancy score is weighted and calculated with the historical sales conversion rate to obtain the comprehensive recommendation score for the compliant recommended object.

8. The sales management method for electrical drive equipment based on knowledge graphs according to claim 4, characterized in that, The method also includes: When none of the candidate targets are identified as compliant recommendation targets, all candidate targets that are permanently blocked are diverted to the manual review queue. Cluster analysis is performed on the reasons for the absence of each candidate target in the manual review queue to generate an early warning report; Obtain feedback operations from the manual review queue, and update the entity relationships in the association model based on the feedback operations, or adjust the parsing rule parameters of the constraint dimension requirements.

9. The sales management method for electrical drive equipment based on knowledge graphs according to claim 1, characterized in that, Before determining compliant recommendation objects from the candidate targets based on the synchronous verification records, preset matching status, and preset performance thresholds, the method further includes: Identify dimension pairs that are physically or logically coupled among the multiple constraint dimension requirements, and use them as a collaborative verification dimension group; The interaction interference coefficients of the candidate targets under the collaborative verification dimension group are extracted from the association model. The interaction interference coefficients are used to characterize the performance loss when different dimensional requirements are implemented simultaneously on the same hardware carrier. Based on the interaction interference coefficient, the corresponding local matching state in the synchronization verification record is reduced to generate a collaborative correction record; The step of determining compliant recommendation objects from the candidate targets based on the synchronous verification records, preset matching states, and preset performance thresholds includes: The candidate target is jointly determined based on the synchronous verification record and the collaborative correction record. The candidate target is determined to be a compliant recommendation object if and only if the candidate target is still in the matching state after considering the interference between dimensions and the performance threshold is met.

10. A knowledge graph-based sales management system for electrical drive equipment, used to execute the steps of the method according to any one of claims 1 to 9, characterized in that, include: The requirement acquisition module is used to acquire multiple constraint dimension requirements of electrical drive equipment and configure a unified association identifier for each constraint dimension requirement; The retrieval record module is used to retrieve candidate targets that match the requirements of each constraint dimension based on a preset association model, and record the local matching status between each candidate target and each constraint dimension requirement. A record construction module is used to assign corresponding local matching states to each candidate target based on the association identifier, so as to generate a synchronization verification record; The joint determination module is used to determine compliant recommended objects from the candidate targets based on the synchronous verification records, preset matching status, and preset performance thresholds.