Hierarchical linkage-based instrument after-sales automation management and quality closed-loop tracing method and system and storage medium

Abstract

This invention discloses a method, system, and storage medium for automated management and quality closed-loop traceability of instrument after-sales service based on hierarchical linkage. The method includes: acquiring target after-sales problem records; constructing a target mapping table based on these records, the target mapping table including data from one or more production workstations corresponding to the target after-sales problem; determining the impact level of the target after-sales problem based on the production workstation data; acquiring the target handling strategy corresponding to the impact level, the target handling strategy including target task scheduling rules and target early warning rules; selecting at least one corresponding notification channel from multiple preset notification channels based on the target early warning rules for early warning push; generating one or more pending tasks based on the production workstation data and the target task scheduling rules and pushing them to one or more target responsible entities corresponding to each pending task. This invention realizes the transformation of instrument after-sales problem handling from manual processing to automated scheduling, improving processing efficiency.

Description

Technical Field

[0001] This invention relates to the field of industrial Internet of Things (IoT), and in particular to a method, system, and storage medium for automated after-sales management and quality closed-loop traceability of instruments based on hierarchical linkage. Background Technology

[0002] In existing technologies, after-sales management of smart water meters, smart electricity meters, and other meter products has long relied on manual work order recording and hierarchical reporting. The current industry practice is as follows: customer service personnel manually enter problem information after receiving customer feedback; after-sales engineers assess the severity of the problem based on experience; and then report it level by level to the relevant departments for processing. Regarding quality traceability, only the product number can be used to query factory inspection records, making it impossible to accurately pinpoint the specific production station.

[0003] The problems with this approach are: after-sales issues cannot be automatically linked to specific production workstations or other production elements; the determination of issue levels and the allocation of handling tasks rely entirely on manual experience, lacking unified standards, leading to untimely handling of urgent issues; the early warning notification method is singular, and key issues cannot be immediately communicated to the responsible parties; when similar issues recur, there is a lack of automatic identification and root cause tracing capabilities, making it impossible to form a closed loop of quality improvement from after-sales to production. These problems are intertwined, resulting in low after-sales processing efficiency, difficulty in quality traceability, and recurring issues.

[0004] Therefore, there is an urgent need for an automated management and quality closed-loop traceability method for instrument after-sales service based on hierarchical linkage, so as to realize instrument after-sales management services with automatic problem classification and closed-loop traceability.

[0005] Therefore, existing technologies still need to be improved and enhanced. Summary of the Invention

[0006] This invention provides a method, system, and storage medium for automated after-sales management and quality closed-loop traceability of instruments based on hierarchical linkage, aiming to solve the problems of low efficiency in after-sales processing, difficulty in quality traceability, and recurring problems in the prior art.

[0007] A first aspect of the present invention provides a method for automated after-sales management and quality closed-loop traceability of instruments based on hierarchical linkage, comprising: Obtain the target after-sales problem record, and construct a target mapping table based on the target after-sales problem record. The target mapping table includes one or more production workstation data corresponding to the target after-sales problem. The impact level of the target after-sales problem is determined based on the production station data. Obtain the target handling strategy corresponding to the level of impact, wherein the target handling strategy includes target task scheduling rules and target early warning rules; Based on the target early warning rule, at least one corresponding notification channel is selected from multiple preset notification channels to push the early warning; Based on the production workstation data and the target task scheduling rules, one or more tasks to be processed are generated and pushed to one or more target responsible entities corresponding to each task to be processed. The receiving status of one or more target responsible entities is continuously tracked. When the task to be processed does not reach the preset processing status within a preset time limit, the task to be processed is automatically upgraded to a higher level of responsible entity. Real-time statistics on the frequency and distribution characteristics of similar after-sales issues; when the statistical results reach the preset non-accidental determination trigger condition, root cause analysis is automatically initiated to output the target rectification instructions. The system obtains feedback results of the target rectification instructions in real time and dynamically optimizes the quality control judgment logic from the sales backend to the production end based on the feedback results.

[0008] A second aspect of the present invention provides an automated after-sales management and quality closed-loop traceability system for instruments based on hierarchical linkage, comprising: A mapping table construction module is used to obtain target after-sales problem records and construct a target mapping table based on the target after-sales problem records. The target mapping table includes one or more production station data corresponding to the target after-sales problem. The determination module is used to determine the impact level of the target after-sales problem based on the production station data; The side strategy acquisition module is used to acquire the target handling strategy corresponding to the impact level, wherein the target handling strategy includes target task scheduling rules and target early warning rules; The push module is used to select at least one corresponding notification channel from multiple preset notification channels to push the warning based on the target warning rule; The processing module is used to generate one or more pending tasks based on the production workstation data and the target task scheduling rules, and push them to one or more target responsible entities corresponding to each pending task. It also continuously tracks the receiving status of one or more target responsible entities. When the pending task does not reach the preset processing status within a preset time limit, it automatically upgrades the pending task to a higher level of responsible entity. The source tracing module is used to statistically analyze the frequency and distribution characteristics of similar after-sales problems in real time. When the statistical results reach the preset non-accidental determination trigger condition, the root cause tracing analysis is automatically started to output the target rectification instructions. The optimization module is used to obtain the feedback results of the target rectification instructions in real time, and dynamically optimize the quality control judgment logic from the back-end to the production end based on the feedback results.

[0009] A third aspect of the present invention provides a storage medium storing one or more programs that can be executed by one or more processors to implement the steps of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage described in any of the preceding claims.

[0010] Beneficial Effects: Compared with existing technologies, this invention provides a method, system, and storage medium for automated after-sales management and quality closed-loop traceability of instruments based on hierarchical linkage. The method for automated after-sales management and quality closed-loop traceability of instruments based on hierarchical linkage provided by this invention involves acquiring target after-sales problem records, constructing a target mapping table based on these records, and the target mapping table including one or more production workstation data corresponding to the target after-sales problem. Then, the impact level of the target after-sales problem is determined based on the production workstation data. Subsequently, a target handling strategy corresponding to the impact level is acquired, and the target handling strategy includes target task scheduling rules and target early warning rules. Based on the target early warning rules, at least one corresponding notification channel is selected from multiple preset notification channels for early warning push. One or more pending tasks are generated based on the production workstation data and the target task scheduling rules and pushed to one or more target responsible entities corresponding to each pending task. The receiving status of one or more target responsible entities is continuously tracked. When a pending task fails to reach a preset processing status within a preset time limit, the pending task is automatically upgraded to a higher-level responsible entity. Following this, the frequency and distribution characteristics of similar after-sales issues will be statistically analyzed in real time. When the statistical results reach the preset non-accidental determination trigger condition, the root cause analysis will be automatically started to output the target rectification instruction, and the feedback results of the target rectification instruction will be obtained in real time. Based on the feedback results, the quality control judgment logic from the after-sales department to the production end will be dynamically optimized. This invention provides users with a hierarchical linkage-based automated management and quality closed-loop traceability method for instrument after-sales service. It achieves this by constructing a mapping between after-sales issues and production workstation data; automatically classifying the impact level of issues based on a dual-mode judgment mechanism and matching corresponding hierarchical handling strategies; providing multi-channel differentiated early warning pushes based on early warning rules; generating and tracking pending tasks according to task scheduling rules until the loop is closed; performing multi-stage root cause analysis on issues that meet non-accidental triggering conditions based on statistical models and outputting rectification instructions; and feeding back and archiving the processing and traceability results while dynamically optimizing the judgment logic. This solves the technical problems in existing instrument after-sales management, such as the disconnect between after-sales and production data, reliance on manual problem handling and lack of a hierarchical linkage mechanism, single early warning channels, and difficulty in accurately tracing the source and achieving closed-loop improvement of quality issues. It achieves the technical effect of full-process quality closed-loop traceability and automated management from the after-sales end to the production end. Attached Figure Description

[0011] Figure 1 A flowchart illustrating an embodiment of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided by the present invention; Figure 2 The overall system architecture diagram is shown in an embodiment of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided by the present invention. Figure 3 A data association mapping diagram is provided in an embodiment of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided by the present invention. Figure 4 This is a schematic diagram of a problem-level decision tree for an embodiment of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided by the present invention. Figure 5 The following is a sequence diagram of the automated processing flow in an embodiment of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided by the present invention. Figure 6 A flowchart illustrating the multi-channel early warning triggering process in an embodiment of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided by the present invention. Figure 7 A flowchart of the quality problem tracing algorithm for an embodiment of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided by the present invention; Figure 8 This is a schematic diagram illustrating the structural principle of an embodiment of the instrument after-sales automated management and quality closed-loop traceability system based on hierarchical linkage provided by the present invention. Detailed Implementation To make the objectives, technical solutions, and effects of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0012] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0013] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0014] Example 1 The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided in this embodiment can be executed by a storage medium, which can be, but is not limited to, a smart computer, a smart mobile device, etc. The following explanation uses smart computing as an example.

[0015] In existing technologies, smart meters, as core terminal devices of IoT technology in the municipal public utility sector, represented by smart water meters and smart electricity meters, undertake multiple functions such as metering, data collection, remote meter reading, and anomaly alarms. In recent years, with the in-depth advancement of smart city and smart energy construction, the shipment volume of smart meters has continued to grow, gradually replacing traditional mechanical meters as the mainstream metering equipment. However, in line with the rapid growth of the market, the after-sales management system of the metering industry as a whole is still in a rough management stage dominated by manual experience.

[0016] After-sales issues with smart meters are characterized by batch-based problems, hidden nature, and delayed response. As mandatory verification measuring instruments, their performance indicators, such as measurement accuracy, sealing, and electronic module stability, are strictly regulated. Batch-based problems can lead to metrological disputes. Furthermore, meters are typically installed in inaccessible locations such as stairwells and underground utility tunnels, resulting in high costs and long service cycles for on-site after-sales support, requiring timely and accurate after-sales dispatch. In addition, the meter manufacturing process involves dozens of steps, including injection molding, machining, assembly, calibration, verification, and testing, spanning multiple factories, workstations, and batches. After-sales problems are often not caused by a single workstation but are the result of multiple processes failing collaboratively, making it difficult to pinpoint the true root cause using traditional single-point tracing methods. Against this backdrop, how to utilize big data, artificial intelligence, and other technologies to build an integrated management system covering the entire process of after-sales problem collection, classification, dispatch, processing, early warning, and traceability has become a core technical challenge that the industry urgently needs to address.

[0017] In existing technologies, after-sales management systems generally suffer from the following defects.

[0018] Firstly, regarding data association, most instrument manufacturers currently use a work order system for recording after-sales issues. Customer service personnel manually enter problem information, and there is a lack of data interfaces between the work order system and the production execution system and quality management system, resulting in information silos between after-sales data and production data. After-sales issue records only contain the product's serial number and cannot be directly linked to key production data such as the factory, production line, workstation, equipment, batch, and operator at the time of production. Problem tracing requires manual cross-system queries, which is extremely inefficient. Different customer service personnel use inconsistent terminology to describe similar problems, making accurate statistical analysis and pattern recognition impossible. After a work order is closed, the data is not fed back into the production quality management system, failing to drive quality improvement, and the same problem recurs in subsequent batches.

[0019] Secondly, at the issue handling process level, traditional after-sales handling relies on manual experience-based judgment and hierarchical reporting, lacking standardized and automated workflow rules. Different personnel may have significantly different judgments on the same issue. The lack of unified level determination standards means that urgent issues may be handled with low priority, leading to escalated customer complaints or increased economic losses. The lack of clear workstation-level responsibility division when issues circulate between departments easily leads to shirking of responsibility. The long manual reporting chain means that urgent issues may take hours or even days from feedback to response, failing to meet service level agreement requirements. Furthermore, the lack of complete operation logs and timestamps during the handling process makes post-event review difficult.

[0020] Thirdly, in terms of quality traceability, existing traceability methods are relatively crude. Most manufacturers can only trace back to a single point by querying the factory inspection record and the final assembly station through the factory serial number. For complex quality issues involving multiple processes and batches, it is difficult to accurately locate the true root cause. It is impossible to distinguish between accidental events and systemic problems, which may lead to over-rectification or missed judgments; it can only locate to the last process before leaving the factory and cannot trace back to upstream processes such as injection molding and machining; no problem threshold is set, so it is impossible to automatically determine whether the batch or systemic standards are met to trigger in-depth traceability; the traceability results are not formed into a knowledge base, and each new problem needs to be re-analyzed, making it impossible to reuse historical experience.

[0021] Fourth, regarding the early warning mechanism, the existing early warning system is primarily reactive, relying on regular manual inspection reports or customer complaint statistics. Only after a certain number of complaints are received are quality management personnel manually triggering the system. Notification methods mainly consist of emails and instant messaging, lacking a differentiated mechanism based on tiers and channels. This results in delayed warnings, with problems accumulating to a certain extent before being discovered, missing the optimal intervention window; a single channel, failing to match differentiated notification methods according to severity; unclear triggering conditions, lacking quantifiable thresholds and multi-dimensional combination judgment rules, easily leading to false alarms or missed alarms; and the lack of real-time tracking of early warning processing progress, posing a risk of unresponsiveness.

[0022] It is evident that the core issue in after-sales management within the instrumentation industry lies in the lack of data correlation, process linkage, and closed-loop feedback mechanisms among the three major systems: after-sales, production, and quality. At the data level, after-sales data cannot be accurately linked to production workstation data; at the process level, problem severity assessment and handling scheduling rely on manual experience; at the early warning level, there is a lack of differentiated and automated multi-channel notifications; and at the traceability level, there is a lack of threshold-based identification capabilities for non-accidental problems. These intertwined issues lead to low after-sales processing efficiency, a lack of data-driven quality improvement, and the recurrence of the same problems, necessitating systemic technological innovation to address them.

[0023] To address the aforementioned shortcomings, this embodiment provides a hierarchical linkage-based automated management and quality closed-loop traceability method for instrument after-sales service. This method aims to automatically link after-sales issues with production workstation data, ensuring that each after-sales issue can be traced back to its corresponding production batch, workstation, equipment, and operator. It also enables automatic determination of the issue's impact, automatic generation of processing tasks, differentiated push notifications, and automatic escalation of timeouts. Furthermore, it provides multi-channel differentiated early warnings based on issue severity, ensuring timely notification and status tracking for critical issues. Finally, it enables automatic identification of non-accidental issues and precise root cause location at the workstation, dynamically optimizing the judgment logic based on traceability results to form a complete quality control closed loop from after-sales to production.

[0024] Specifically, such as Figure 1 As shown, the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage provided in this embodiment includes the following steps: S100. Obtain the target after-sales problem record, and construct a target mapping table based on the target after-sales problem record. The target mapping table includes one or more production station data corresponding to the target after-sales problem.

[0025] Reference Figure 2 , Figure 2 This is a diagram of the overall system architecture. In this embodiment, based on... Figure 2The system shown in the following implements the instrument after-sales automated management and quality closed-loop tracing method based on hierarchical linkage. The system adopts a four-layer architecture, which from bottom to top are respectively a data acquisition layer, a data association layer, a business logic layer and an application presentation layer. The two closed loops are respectively an after-sales processing closed loop and a quality improvement closed loop, which share the same set of data base through the data association layer to realize linkage and collaboration. The data acquisition layer is responsible for collecting three types of data including after-sales data, production data and quality data from multi-source heterogeneous systems, and completes data cleaning, field alignment and format standardization through a unified data access gateway. The data sources include: after-sales side - customer service work order system, remote water meter uploaded data, customer APP / miniprogram feedback, on-site service engineer mobile terminal report; production side - station production records, equipment operation parameters, batch material information, operator operation logs of the MES system (Manufacturing Execution System); quality side - factory inspection records, SPC (Statistical Process Control) process quality data, verification / calibration data of the QMS system (Quality Management System). The data access gateway supports four access modes: RESTful API (Representational State Transfer Application Programming Interface), message queue, batch file import, and MQTT (Message Queuing Telemetry Transport) direct connection with the IoT (Internet of Things) platform. After the data acquisition layer completes data aggregation, the data association layer undertakes the function of precise docking of after-sales data and production data, and this layer is built-in with unified data standards, an association rule engine, a mapping relationship table and a master data management module.

[0026] In this embodiment, constructing a target mapping table based on the target after-sales problem record includes: Taking the unique identity identifier of the target instrument corresponding to the target after-sales problem record as an index, associating it to the target production batch and final assembly station corresponding to the target instrument; Acquiring all stations that the target production batch has passed through, all used equipment and operators, and generating a mapping relationship between the target after-sales problem and one or more pieces of production station data; Each mapping record in the target mapping table includes a confidence score, which is calculated as the weighted sum of the proportion of historical similar problems, the matching weight between the station and the problem type, and the time decay factor.

[0027] Specifically, in this embodiment, after-sales problem data is stored and managed in the form of an after-sales problem data table (denoted as T_AFTER_SALES_ISSUE). This data table is shown in Table 1.

[0028] Table 1:

[0029] Production station data is stored in the form of a production station data table (denoted as T_PROD_STATION). This data table is shown in Table 2.

[0030] Table 2:

[0031] Reference Figure 3 , Figure 3 This diagram illustrates the mapping relationship between after-sales issues and production workstation data. In this embodiment, the association rule engine uses the instrument's unique identification code, i.e., the instrument's factory serial number, as the primary key. Through preset association mapping rules, it establishes a one-to-many relationship between each after-sales issue record and key production elements in the instrument's production process, such as the factory, production line, workstation, equipment, batch, and operator. The establishment of the relationship adopts a two-level mechanism of "primary key association + element expansion": the first level associates the instrument's factory serial number with the instrument's production batch and final assembly workstation; the second level associates the batch number with all workstations, all equipment and materials used, and all operators involved in that batch, forming a complete "after-sales issue - production element" association graph.

[0032] The mapping relationship is established using an association mapping table (denoted as T_ISSUE_STATION_MAP). In this embodiment, the association mapping table corresponding to the target after-sales issue record is the target mapping table, and the target after-sales issue can be one or more. The target mapping table serves as a bridge table between the target after-sales issue and the production workstation, recording the many-to-many relationships between the target after-sales issue and multiple associated production elements. Specifically, this mapping table uses meter_sn as an indirect association key to establish associations between the after-sales issue data table and the production-side master data tables such as the production workstation data table, production batch data table, production equipment data table, and production operator data table.

[0033] Each mapping record in the target mapping table includes a confidence score, which is used to rank candidate root cause workstations during subsequent source tracing analysis. After the association is established, the system automatically generates a confidence score for each mapping record. The confidence score is calculated based on the weighted sum of three factors: the proportion of similar problems in history, the matching weight between the workstation and the problem type, and the time decay factor. Candidate root cause workstations are ranked from highest to lowest confidence score for use in subsequent source tracing analysis.

[0034] Based on this, the time for related queries has been reduced from several hours to seconds.

[0035] S200: Determine the impact level of the target after-sales problem based on the production station data; Based on the production station data, the impact level of the target after-sales issue is determined, including: The target after-sales problem is classified and determined in parallel using a preset decision tree algorithm and a preset rule engine. When the number of historical labeled samples exceeds a preset threshold, the influence level is determined based on the decision tree algorithm, and the determination path and the contribution of each feature are output. When the number of historical labeled samples is less than or equal to the preset threshold, the influence level is determined based on the rule engine.

[0036] like Figure 4 As shown in the figure, this embodiment classifies after-sales issues into four levels: L1 Emergency, L2 Severe, L3 General, and L4 Minor, with the level decreasing from high to low. The determination is based on a comprehensive consideration of the impact of the issue on the instrument's metering function, safety, user experience, and compliance. The issue level determination criteria are shown in Table 3.

[0037] Table 3:

[0038] Specifically, the L1 emergency level corresponds to situations such as serious measurement deviation (more than 10%) or complete failure, water leakage posing a risk to personal injury or property, batches (10 or more units) being unusable, and violations of mandatory verification procedures. Typical problems include batch water leakage and measurement deviation exceeding 10%. The L2 severe level corresponds to measurement deviation (5% to 10%), potential safety hazards, single unit being unusable or batches (3 to 9 units) malfunctioning, and approaching compliance thresholds. Typical problems include electronic module failure and display abnormalities. The L3 general level corresponds to measurement deviation (2% to 5%), no safety hazards, single unit malfunctioning but not affecting the main function, and not affecting compliance. Typical problems include blurry display and button malfunction. The L4 minor level corresponds to measurement deviation not exceeding 2% or no impact, no safety impact, appearance defects or slight noise, and not affecting compliance. Typical problems include appearance scratches and unclear markings.

[0039] In this embodiment, the judgment mechanism adopts a dual judgment mode of a preset decision tree algorithm and a preset rule engine, which are used in parallel to classify and determine the target after-sales problem. The rule engine is responsible for making a quick initial judgment based on structured fields, such as problem category, scope of impact, and batch size; the decision tree algorithm is responsible for making a secondary fine judgment by combining unstructured fields, such as problem description text, and finally outputting the level judgment result and judgment basis.

[0040] In this embodiment, the decision tree algorithm takes five dimensions as input features: problem classification, scope of impact, batch size, customer type, and compliance risk, and outputs the problem level. The decision tree is trained based on the company's historical after-sales data, i.e., labeled samples, and is constructed using the C4.5 algorithm, supporting online incremental learning. The decision tree's decision process is interpretable; each decision outputs the decision path and the contribution of each feature, facilitating manual review by quality engineers and model iteration.

[0041] When the number of historical labeled samples exceeds the preset threshold, the impact level is determined based on a decision tree algorithm, and the determination path and the contribution of each feature are output. When the number of historical labeled samples is less than or equal to the preset threshold, the impact level is determined based on a rule engine. That is, for new problem types lacking historical data, the system reverts to the rule engine for threshold-based determination.

[0042] The preset threshold is used to determine whether there are enough labeled samples to train / use the decision tree model. According to conventional practice in the field of machine learning, the preset threshold is usually in the range of 100 to 500, and in this embodiment, it can be 200.

[0043] Furthermore, in this embodiment, in addition to the dual-mode judgment, a conservative principle and special escalation rules are introduced. Specifically, when the judgment results of the decision tree algorithm and the rule engine are inconsistent, the system follows the conservative principle and adopts the higher level of the two as the final output. At the same time, the system is also configured with special escalation rules, including but not limited to: automatic escalation when the number of similar issues in the same batch reaches the batch size threshold, and automatic escalation when the issue involves key customers.

[0044] When the problem level judgment result is inconsistent with the manual review result, the system automatically adds the sample to the incremental learning queue, retrains the decision tree model regularly, and continuously improves the judgment accuracy.

[0045] S300. Obtain the target handling strategy corresponding to the level of influence, wherein the target handling strategy includes target task scheduling rules and target early warning rules; Obtaining the target handling strategy corresponding to the level of impact includes: Obtain a preset level-policy mapping relationship, and match the target task scheduling rule and the target early warning rule corresponding to the level of influence based on the level-policy mapping relationship. Different target task scheduling rules include different service response time limits, service processing time limits and responsibility flow paths, and the service response time limit and the service processing time limit are gradually relaxed as the level of influence decreases.

[0046] In this embodiment, the system has a preset level-policy mapping relationship. Based on the level of influence, the corresponding target task scheduling rules and target early warning rules are matched according to the level-policy mapping relationship. Specifically, the target handling strategy includes two dimensions: one is the task scheduling rules, which guide the generation, push, transfer, and escalation of tasks to be processed; the other is the early warning rules, which guide the determination of the early warning level and the selection of the early warning channel.

[0047] The target task scheduling rules are configured with differentiated service response time limits, service processing time limits, and responsibility transfer paths based on different problem levels. The service response time limit refers to the allowed time from task push to the responsible person and to the responsible person's confirmation of receipt; the service processing time limit refers to the allowed time from task confirmation to task completion. The service response time limit and service processing time limit are progressively relaxed as the impact level decreases.

[0048] Specifically, Table 4 is the rule table for generating pending tasks.

[0049] Table 4:

[0050] Specifically, for the L1 emergency level, the root cause workstation and its upstream and downstream workstations are associated with the task, which involves emergency rectification plus root cause analysis. The responsible roles are workstation supervisor, workshop director, and quality manager. The service response time is 15 minutes, and the service processing time is 4 hours. The responsibility flow path escalates sequentially from workstation to workshop to quality to factory manager. For the L2 severe level, the root cause workstation and adjacent workstations are associated with the task, which involves rectification plus root cause analysis. The responsible roles are workstation supervisor and workshop director. The service response time is 1 hour, and the service processing time is 24 hours. The responsibility flow path escalates sequentially from workstation to workshop to quality. For the L3 general level, the root cause workstation is associated with the task, which involves rectification. The responsible roles are workstation operator and workstation supervisor. The service response time is 4 hours, and the service processing time is 72 hours. The responsibility flow path escalates sequentially from workstation to workshop. For L4 minor level, the associated root cause workstation, the task type is record and periodic rectification, the responsible role is the workstation operator, the service response time is 24 hours, the service processing time is 7 days, and the responsibility transfer path only involves the workstation level.

[0051] In this embodiment, each pending task has six states: pending receipt → received → processing → processed → verified → closed loop, driven by the workflow engine. If a task remains in a certain state for longer than the SLA limit, the system automatically triggers an escalation, transferring the task to the next higher-level responsible role and simultaneously triggering a corresponding level of alert. Each state change records the operator, operation time, and operation content, forming a complete processing audit log.

[0052] like Figure 5 As shown, the specific execution sequence of the target task scheduling rule is as follows: After the after-sales issue is entered into the database (T0), the data association layer completes the association of production elements (T0+10 seconds), the level determination engine outputs the issue level (T0+15 seconds), and the task scheduling engine generates a pending task based on the level and associated workstation and pushes it to the responsible person (T0+30 seconds); the responsible person receives the task and starts processing, and the system starts the SLA timer; during the processing, photos can be uploaded, processing records can be filled in, and the knowledge base can be called; after processing is completed, the submission is verified, and the processing result is verified by the quality engineer or customer service personnel; after verification, the task is closed, the issue data is automatically archived to the quality knowledge base and the source analysis is triggered; if the source analysis identifies it as a non-accidental issue, the in-depth rectification process is automatically triggered.

[0053] S400. Based on the target warning rule, select at least one corresponding notification channel from multiple preset notification channels to push the warning. Based on the aforementioned warning rules, at least one corresponding notification channel is selected from multiple preset notification channels to push the warning, including: The target warning level is determined according to the aforementioned warning rules; When the target warning level is the highest level, at least an automatic voice notification channel shall be used to push the warning to the main responsible party, and the recipient's answering confirmation status shall be obtained through voice interaction. If the answer confirmation status is not obtained within the preset time, a backup responsible party is obtained and automatically transferred to the backup responsible party to send an early warning push to the backup responsible party, and at the same time, a supplementary SMS notification is sent to the main responsible party; During the processing of the pending task, the processing progress is checked back through an automatic voice notification channel at a preset frequency until the pending task is completed.

[0054] like Figure 6 As shown, in this embodiment, the target warning level is determined according to the target warning rules. This embodiment configures a four-level warning mechanism for different levels of quality problems. The warning levels are divided into P0 (Special), P1 (High), P2 (Medium), and P3 (Low) from high to low, as detailed in Table 5.

[0055] Table 5:

[0056] Specifically, when the target warning level is the highest level, i.e., P0 (Special Level), corresponding to a batch outbreak of L1 emergency issues or non-accidental issues, the system will at least use an automatic voice notification channel to push warnings to the main responsible parties. The automatic voice notification adopts an AI outbound calling solution based on speech synthesis and speech recognition technology to achieve second-level reach for critical issues. Specifically, after the warning engine triggers a P0 warning, the system calls the AI ​​outbound calling API to generate an outbound calling task based on the warning level and the responsible person's schedule information. The AI ​​outbound calling platform automatically dials the responsible person's phone, plays the warning content including a problem summary, associated workstation, processing requirements, and SLA deadline, and obtains the recipient's answer confirmation status through voice interaction (button confirmation). The responsible person confirms receipt by pressing a button, and the system records the answer time and confirmation status.

[0057] If no answer confirmation is received within the preset time, a backup responsible party is automatically identified and the call is forwarded to that party to receive a warning notification. Simultaneously, a supplementary SMS notification is sent to the primary responsible party. In other words, the call is automatically forwarded to the next level of responsible person, and a backup SMS notification is sent. The preset time represents the window duration during which the system waits for answer confirmation after the automatic voice notification channel dials the responsible party's number. In this embodiment, the preset time can be 30 seconds.

[0058] During the processing of pending tasks, the system follows up on the processing progress via automated voice notification at a preset frequency. Specifically, after an alert is issued, the system periodically calls back to check on the processing progress until the pending task is completed, thus achieving full tracking of the alert processing progress. The preset frequency represents the interval at which the system periodically checks on the processing progress via automated voice notification after the alert is issued and before the task is closed. In this embodiment, the preset interval can be 2 hours.

[0059] For warning levels other than the highest level, this embodiment also configures differentiated notification channel combinations: P1 high corresponds to L2 serious issues or three consecutive similar issues at a single workstation, using AI phone calls and SMS channels; P2 medium corresponds to L3 general issues or monthly issues exceeding the threshold at a single workstation, using SMS and system pop-up channels; P3 low corresponds to L4 minor issues or periodic quality reports, using email and system pop-up channels.

[0060] As can be seen, in this embodiment, the highest level of warning can reach the responsible person within 30 seconds, and the warning response rate is over 99%.

[0061] S500: Based on the production workstation data and the target task scheduling rules, generate one or more tasks to be processed and push them to one or more target responsible entities corresponding to each task to be processed, and continuously track the receiving status of one or more target responsible entities. When the task to be processed does not reach the preset processing status within a preset time limit, automatically upgrade the task to a higher level of responsible entity. Based on the production workstation data and the task scheduling rules, one or more tasks to be processed are generated and pushed to one or more target responsible entities corresponding to each task to be processed, including: The preset response time limit and preset processing time limit for each task to be processed, as well as the preset flow order of each task to be processed, are determined according to the task scheduling rules. According to the preset flow order, each task to be processed is pushed to the corresponding target responsible entity. The workflow engine drives the flow status of each task to be processed to go through six states in sequence: waiting to be received, received, processing, processed, verified, and closed loop. During the workflow of the tasks to be processed, the operator, operation time, and operation content of each status change are recorded to form an audit log.

[0062] In this embodiment, firstly, based on the task scheduling rules, the preset response time limit and preset processing time limit for each task to be processed, as well as the preset flow order of each task to be processed, are determined. The preset response time limit refers to the allowed time from when the task is pushed to the responsible entity until the responsible entity confirms receipt, corresponding to the "SLA Response Time Limit" in Table 2; the preset processing time limit refers to the allowed time from when the responsible entity confirms receipt until the task is completed, corresponding to the "SLA Processing Time Limit" in Table 2. Specifically, the preset response time limit for the L1 emergency level is 15 minutes, the preset processing time limit is 4 hours, and the responsibility transfer path is escalated sequentially from the workstation to the workshop to quality control to the factory manager; the preset response time limit for the L2 severe level is 1 hour, the preset processing time limit is 24 hours, and the responsibility transfer path is escalated sequentially from the workstation to the workshop to quality control; the preset response time limit for the L3 general level is 4 hours, the preset processing time limit is 72 hours, and the responsibility transfer path is escalated sequentially from the workstation to the workshop; the preset response time limit for the L4 minor level is 24 hours, the preset processing time limit is 7 days, and the responsibility transfer path only involves the workstation level.

[0063] Then, each pending task is pushed to the corresponding responsible entity according to the preset workflow sequence. The workflow engine drives the workflow status of each pending task through six states in sequence: pending receipt, received, processing, processed, verified, and closed loop. Specifically, "pending receipt" indicates that the task has been generated and pushed to the responsible entity, awaiting confirmation of receipt; "received" indicates that the responsible entity has confirmed receipt of the task and has entered the processing preparation stage; "processing" indicates that the responsible entity is performing processing operations; "processed" indicates that the responsible entity has completed the processing operations and submitted the processing results for verification; "verified" indicates that the quality engineer or customer service personnel have verified the processing results and the verification has passed; and "closed loop" indicates that the problem has been completely resolved and the task is officially closed.

[0064] Finally, during the workflow of the pending tasks, the operator, time, and content of each status change are recorded to form a complete audit log, enabling full-chain traceability of the processing. If a task remains in a certain status for longer than the preset SLA (Service Level Agreement) time limit, the system automatically triggers an escalation, transferring the task to the next higher-level responsible party. Specifically, when a task remains in the "Pending Receive" status for longer than the preset response time limit, the system automatically escalates to a higher-level responsible entity; when a task remains in the "Received" or "Processing" status for longer than the preset processing time limit, the system also automatically escalates and simultaneously triggers the corresponding level of warning, ensuring timely task processing.

[0065] S600: Real-time statistics on the frequency and distribution characteristics of similar after-sales issues. When the statistical results reach the preset non-accidental determination trigger condition, the root cause analysis is automatically started to output the target rectification instructions. The root cause analysis includes: Obtain historical statistical data, and based on the historical statistical data, select at least one target candidate workstation from multiple candidate production workstations whose problem contribution meets preset conditions; Based on a preset statistical process control model, it is determined whether the problem occurrence rate of the target candidate workstation is out of control; When it is determined that the problem occurrence rate of the target candidate workstation is out of control, the posterior probability that the target candidate workstation is the root cause of the problem is calculated based on Bayesian inference. Based on the comparison between the posterior probability and the preset probability threshold, the root cause localization result and the corresponding rectification instructions are output.

[0066] like Figure 7As shown, in this embodiment, the frequency and distribution characteristics of similar after-sales problems are statistically analyzed in real time. In this embodiment, three types of non-accidental problem identification thresholds are set as trigger conditions: batch size threshold (3 or more units with the same type of problem in the same batch are judged as batch-related non-accidental problems); workstation frequency threshold (5 or more instances of the same type of problem in the same workstation within 30 days are judged as workstation-related non-accidental problems); and control chart threshold (workstation problem incidence rate exceeds the control limit, judged as systemic non-accidental problems). Triggering any of these three thresholds automatically initiates the root cause analysis.

[0067] The root cause analysis employs a three-stage progressive algorithm, including the Pareto Analysis stage, the control chart stage, and the Bayesian inference stage.

[0068] Specifically, the first stage is the Pareto analysis stage. The system acquires historical statistical data and, based on this data, analyzes the same type of problem within the past 30 days by workstation. From multiple candidate production workstations, it selects at least one target candidate workstation whose problem contribution meets preset conditions. Specifically, Pareto analysis identifies the top 3 workstations with the highest contribution as candidate root cause workstations. The threshold for inclusion is a single workstation's contribution ≥ 20%, which is then added to the candidate root cause workstation list. Pareto analysis outputs a workstation-problem Pareto chart, allowing quality engineers to intuitively identify the critical few workstations.

[0069] The second stage is the control chart stage. The system determines whether the problem incidence rate of the target candidate workstation is out of control based on a preset statistical process control model. Specifically, SPC analysis is performed on the historical problem incidence rate of the candidate root cause workstation, and a p-chart (nonconforming rate control chart) is used to identify whether the workstation is in statistical control. The control chart calculates the upper and lower control limits based on data from the past 90 production days of the workstation. In this embodiment, the formula for calculating the upper control limit (UCL) is: UCL = p_bar + 3 × sqrt(p_bar × (1-p_bar) / n), where p_bar is the historical average nonconforming rate and n is the sample size. If the recent problem incidence rate exceeds the upper control limit, the workstation is determined to be out of control, indicating a systemic quality problem.

[0070] The third stage is the Bayesian inference stage. When the occurrence rate of a problem identified as a target candidate workstation is out of control, the system calculates the posterior probability that the target candidate workstation is the root cause of the problem based on Bayesian inference. Specifically, by combining the prior probability (historical problem distribution) and the likelihood function (current problem characteristics), the system outputs the posterior probability that each candidate workstation is the true root cause.

[0071] Finally, based on the comparison between the posterior probability and the preset probability threshold, the root cause localization result and corresponding rectification instructions are output. The preset probability threshold represents the critical posterior probability value used in the Bayesian inference stage to determine whether a candidate workstation is the true root cause. It is used to quantify the source tracing analysis results into actionable judgment conclusions, enabling automatic identification and hierarchical processing of root cause workstations. In this embodiment, the preset probability threshold includes a first preset probability threshold, i.e., a high-confidence threshold, with a value of 0.7; and a second preset probability threshold, i.e., a suspicious threshold, with a value of 0.5. Specifically, workstations with a posterior probability exceeding the first preset probability threshold are determined to be high-confidence root cause workstations, and the system automatically triggers a deep rectification task and outputs corresponding rectification instructions; workstations with a posterior probability between the second preset probability threshold and the first preset probability threshold are determined to be suspicious root cause workstations, requiring manual review to determine whether to trigger rectification; workstations with a posterior probability below the second preset probability threshold are not temporarily subject to rectification. After the source tracing analysis is completed, the analysis results are pushed to the quality manager and production manager to drive targeted rectification.

[0072] In this embodiment, the accuracy rate of identifying non-accidental problems can reach over 85%, and the recurrence rate of the same problem is reduced by over 60%.

[0073] S700: Obtain feedback results of the target rectification instructions in real time, and dynamically optimize the quality control judgment logic from the sales backend to the production end based on the feedback results.

[0074] Based on the feedback results, the quality control judgment logic from the after-sales service to the production end is dynamically optimized, including: The processing data of the target after-sales issues and the results of the root cause analysis are archived into the target instrument knowledge base; When the cumulative amount of data for the same type of problem reaches the preset retraining threshold, the incremental retraining of the decision tree algorithm is automatically triggered, and the decision rules of the decision tree algorithm are updated. And / or, automatically calibrate the threshold parameters of the non-accidental determination trigger condition.

[0075] Specifically, in this embodiment, the method also includes obtaining feedback results of the target rectification instructions in real time. Specifically, after the target rectification instructions are pushed to the production end, the relevant responsible parties on the production end execute the rectification and, upon completion, feed back the execution results to the system. After receiving the feedback results, the system initiates a dynamic optimization process for the quality control judgment logic.

[0076] First, the system archives the processing data of the target after-sales issues and the results of root cause analysis into the target instrument's knowledge base. Specifically, the entire process data of after-sales issues from entry to closure—including problem description, level determination results, processing records, verification results, root cause analysis conclusions, and corrective measures—is stored in the knowledge base according to a preset data structure, forming a structured historical case library that supports rapid retrieval and reuse of processing experience for similar issues in the future.

[0077] When the cumulative amount of data for the same type of problem reaches a preset retraining threshold, the system automatically triggers incremental retraining of the decision tree algorithm and updates the decision tree algorithm's judgment rules. The preset retraining threshold represents the critical number of labeled samples of the same type of problem required to trigger incremental retraining of the decision tree model. The preset retraining threshold can be between 100 and 500 samples.

[0078] Each closed-loop after-sales issue is input as a feedback signal to the model iteration queue. When the cumulative amount of data for the same type of issue reaches the preset retraining threshold (e.g., accumulating 500 labeled samples), the system automatically triggers incremental retraining of the decision tree model. The newly added labeled samples are used to iteratively optimize the model, update the decision tree's judgment rules and the splitting thresholds of each feature, so that the level judgment accuracy continues to improve as the system runs.

[0079] Meanwhile, the system automatically calibrates the threshold parameters for non-randomness detection trigger conditions. Specifically, the system recalculates the contribution threshold of Pareto analysis, the control upper limit of control charts, and the thresholds for identifying various non-randomness issues based on the latest historical data. This allows the threshold parameters to adapt to dynamic changes such as changes in production processes, raw material replacements, and seasonal factors, avoiding false alarms or missed alarms caused by fixed thresholds.

[0080] Through the above mechanism, the system establishes a complete closed loop of "judgment-processing-verification-feedback". After each after-sales issue is processed, the processing result and the source tracing conclusion are input back into the level judgment model and the source tracing trigger conditions. This allows the system's judgment logic to be continuously optimized as data accumulates, realizing dynamic self-optimization of the quality control judgment logic from the after-sales end to the production end, forming a closed loop of full-process quality control that is "perceptible, traceable, and predictable after-sales".

[0081] In summary, this embodiment provides a hierarchical linkage-based automated management and quality closed-loop traceability method for instrument after-sales service. It acquires target after-sales problem records, constructs a target mapping table based on these records, and includes data from one or more production workstations corresponding to the target after-sales problem. Then, it determines the impact level of the target after-sales problem based on the production workstation data. Next, it acquires the target handling strategy corresponding to the impact level, which includes target task scheduling rules and target early warning rules. Based on the target early warning rules, it selects at least one corresponding notification channel from multiple preset notification channels for early warning push. Furthermore, it generates one or more pending tasks based on the production workstation data and the target task scheduling rules and pushes them to one or more target responsible entities corresponding to each pending task. It continuously tracks the reception status of one or more target responsible entities. When a pending task fails to reach a preset processing status within a preset time limit, it is automatically escalated to a higher-level responsible entity. Following this, the frequency and distribution characteristics of similar after-sales issues will be statistically analyzed in real time. When the statistical results reach the preset non-accidental determination trigger condition, the root cause analysis will be automatically started to output the target rectification instruction, and the feedback results of the target rectification instruction will be obtained in real time. Based on the feedback results, the quality control judgment logic from the after-sales department to the production end will be dynamically optimized. This embodiment provides users with a hierarchical linkage-based automated management and quality closed-loop traceability method for instrument after-sales service. It can construct a mapping between after-sales issues and production workstation data; automatically classify the impact level of issues based on a dual-mode judgment mechanism and match corresponding hierarchical handling strategies; push multi-channel differentiated early warnings according to early warning rules; generate and track pending tasks according to task scheduling rules until the loop is closed; perform multi-stage root cause analysis on issues that meet non-accidental triggering conditions based on statistical models and output rectification instructions; and archive the processing process and traceability results and dynamically optimize the judgment logic. This solves the technical problems in existing instrument after-sales management, such as the disconnect between after-sales data and production data, reliance on manual problem handling and lack of hierarchical linkage mechanisms, single early warning channels, and difficulty in accurately tracing the source and achieving closed-loop improvement of quality issues. It achieves the technical effect of full-process quality closed-loop traceability and automated management from the after-sales end to the production end.

[0082] It should be understood that although the steps in the flowcharts shown in the accompanying drawings are displayed sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of the steps in this invention, and these steps can be executed in other orders. Moreover, at least a portion of the steps in this invention may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0083] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program using signal-related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchlink, DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0084] Example 2 Based on the above embodiments, the present invention also provides an automated management and quality closed-loop traceability system for instrument after-sales service based on hierarchical linkage, the functional module diagram of which is shown below. Figure 8 As shown, this instrument after-sales automated management and quality closed-loop traceability system based on hierarchical linkage includes: The mapping table construction module is used to obtain target after-sales problem records and construct a target mapping table based on the target after-sales problem records. The target mapping table includes one or more production workstation data corresponding to the target after-sales problem, as specifically described in Embodiment 1. The determination module is used to determine the impact level of the target after-sales problem based on the production station data, as specifically described in Embodiment 1; The side strategy acquisition module is used to acquire the target handling strategy corresponding to the influence level. The target handling strategy includes target task scheduling rules and target early warning rules, as described in Embodiment 1. The push module is used to select at least one corresponding notification channel from multiple preset notification channels to push the warning based on the target warning rule, as described in Embodiment 1. The processing module is used to generate one or more pending tasks based on the production workstation data and the target task scheduling rules, and push them to one or more target responsible entities corresponding to each pending task. It also continuously tracks the receiving status of one or more target responsible entities. When the pending task does not reach the preset processing status within a preset time limit, it automatically upgrades the pending task to a higher-level responsible entity, as described in Embodiment 1. The source tracing module is used to statistically analyze the frequency and distribution characteristics of similar after-sales problems in real time. When the statistical results reach the preset non-accidental determination trigger condition, the root cause tracing analysis is automatically started to output the target rectification instructions, as described in Example 1. The optimization module is used to obtain the feedback results of the target rectification instructions in real time, and dynamically optimize the quality control judgment logic from the back-end to the production end based on the feedback results, as described in Embodiment 1.

[0085] Example 3 The present invention also provides a storage medium storing one or more programs, which can be executed by one or more processors to implement the steps of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage described in the above embodiments.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for automated after-sales management and quality closed-loop traceability of instruments based on hierarchical linkage, characterized in that, include: Obtain the target after-sales problem record, and construct a target mapping table based on the target after-sales problem record. The target mapping table includes one or more production workstation data corresponding to the target after-sales problem. The impact level of the target after-sales problem is determined based on the production station data. Obtain the target handling strategy corresponding to the level of impact, wherein the target handling strategy includes target task scheduling rules and target early warning rules; Based on the target early warning rule, at least one corresponding notification channel is selected from multiple preset notification channels to push the early warning; Based on the production workstation data and the target task scheduling rules, one or more tasks to be processed are generated and pushed to one or more target responsible entities corresponding to each task to be processed. The receiving status of one or more target responsible entities is continuously tracked. When the task to be processed does not reach the preset processing status within a preset time limit, the task to be processed is automatically upgraded to a higher level of responsible entity. Real-time statistics on the frequency and distribution characteristics of similar after-sales issues; when the statistical results reach the preset non-accidental determination trigger condition, root cause analysis is automatically initiated to output the target rectification instructions. The system obtains feedback results of the target rectification instructions in real time and dynamically optimizes the quality control judgment logic from the sales backend to the production end based on the feedback results.

2. The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage as described in claim 1, characterized in that, A target mapping table is constructed based on the target after-sales issue records, including: Using the unique identifier of the target instrument corresponding to the target after-sales problem record as an index, it is associated with the target production batch and final assembly station corresponding to the target instrument; Obtain the workstations through which the target production batch passed, all the equipment used, and the operators, and generate a mapping relationship between the target after-sales problem and one or more production workstation data; Each mapping record in the target mapping table contains a confidence score, which is calculated based on the weighted sum of three factors: the proportion of similar problems in history, the matching weight between workstation and problem type, and the time decay factor.

3. The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage as described in claim 1, characterized in that, Based on the production station data, the impact level of the target after-sales issue is determined, including: The target after-sales problem is classified and determined in parallel using a preset decision tree algorithm and a preset rule engine. When the number of historical labeled samples exceeds a preset threshold, the influence level is determined based on the decision tree algorithm, and the determination path and the contribution of each feature are output. When the number of historical labeled samples is less than or equal to the preset threshold, the influence level is determined based on the rule engine.

4. The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage as described in claim 1, characterized in that, Obtaining the target handling strategy corresponding to the level of impact includes: Obtain a preset level-policy mapping relationship, and match the target task scheduling rule and the target early warning rule corresponding to the level of influence based on the level-policy mapping relationship. Different target task scheduling rules include different service response time limits, service processing time limits and responsibility flow paths, and the service response time limit and the service processing time limit are gradually relaxed as the level of influence decreases.

5. The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage according to claim 1, characterized in that, Based on the aforementioned warning rules, at least one corresponding notification channel is selected from multiple preset notification channels to push the warning, including: The target warning level is determined according to the aforementioned warning rules; When the target warning level is the highest level, at least an automatic voice notification channel shall be used to push the warning to the main responsible party, and the recipient's answering confirmation status shall be obtained through voice interaction. If the answer confirmation status is not obtained within the preset time, a backup responsible party is obtained and automatically transferred to the backup responsible party to send an early warning push to the backup responsible party, and at the same time, a supplementary SMS notification is sent to the main responsible party; During the processing of the pending task, the processing progress is checked back through an automatic voice notification channel at a preset frequency until the pending task is completed.

6. The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage according to claim 1, characterized in that, Based on the production workstation data and the task scheduling rules, one or more tasks to be processed are generated and pushed to one or more target responsible entities corresponding to each task to be processed, including: The preset response time limit and preset processing time limit for each task to be processed, as well as the preset flow order of each task to be processed, are determined according to the task scheduling rules. According to the preset flow order, each task to be processed is pushed to the corresponding target responsible entity. The workflow engine drives the flow status of each task to be processed to go through six states in sequence: waiting to be received, received, processing, processed, verified, and closed loop. During the workflow of the tasks to be processed, the operator, operation time, and operation content of each status change are recorded to form an audit log.

7. The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage according to claim 1, characterized in that, The root cause analysis includes: Obtain historical statistical data, and based on the historical statistical data, select at least one target candidate workstation from multiple candidate production workstations whose problem contribution meets preset conditions; Based on a preset statistical process control model, it is determined whether the problem occurrence rate of the target candidate workstation is out of control; When it is determined that the problem occurrence rate of the target candidate workstation is out of control, the posterior probability that the target candidate workstation is the root cause of the problem is calculated based on Bayesian inference. Based on the comparison between the posterior probability and the preset probability threshold, the root cause localization result and the corresponding rectification instructions are output.

8. The instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage according to claim 3, characterized in that, Based on the feedback results, the quality control judgment logic from the after-sales service to the production end is dynamically optimized, including: The processing data of the target after-sales issues and the results of the root cause analysis are archived into the target instrument knowledge base; When the cumulative amount of data for the same type of problem reaches the preset retraining threshold, the incremental retraining of the decision tree algorithm is automatically triggered, and the decision rules of the decision tree algorithm are updated. And / or, automatically calibrate the threshold parameters of the non-accidental determination trigger condition.

9. A hierarchical linkage-based automated management and quality closed-loop traceability system for instrument after-sales service, characterized in that, The system includes: A mapping table construction module is used to obtain target after-sales problem records and construct a target mapping table based on the target after-sales problem records. The target mapping table includes one or more production station data corresponding to the target after-sales problem. The determination module is used to determine the impact level of the target after-sales problem based on the production station data; The side strategy acquisition module is used to acquire the target handling strategy corresponding to the impact level, wherein the target handling strategy includes target task scheduling rules and target early warning rules; The push module is used to select at least one corresponding notification channel from multiple preset notification channels to push the warning based on the target warning rule; The processing module is used to generate one or more pending tasks based on the production workstation data and the target task scheduling rules, and push them to one or more target responsible entities corresponding to each pending task. It also continuously tracks the receiving status of one or more target responsible entities. When the pending task does not reach the preset processing status within a preset time limit, it automatically upgrades the pending task to a higher level of responsible entity. The source tracing module is used to statistically analyze the frequency and distribution characteristics of similar after-sales problems in real time. When the statistical results reach the preset non-accidental determination trigger condition, the root cause tracing analysis is automatically started to output the target rectification instructions. The optimization module is used to obtain the feedback results of the target rectification instructions in real time, and dynamically optimize the quality control judgment logic from the back-end to the production end based on the feedback results.

10. A storage medium, characterized in that, The storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the instrument after-sales automated management and quality closed-loop traceability method based on hierarchical linkage as described in any one of claims 1-8.

Images