New energy automobile pressure sensor production visual management system and method

Abstract

The application relates to the field of pressure sensor production management, and discloses a new energy automobile pressure sensor production visual management system and method, which realizes real-time collection of working parameters, equipment states, product identification, flow data and test result data by deploying an IIoT gateway in a production line, and comprehensively digitizes key information in the production process. Meanwhile, by constructing pressure sensor production graph data, the production data is mapped into a graph database structure, thereby forming a flexible and efficient data relationship management and analysis mechanism. By using the graph data structure, quality alarms can be quickly responded to, potential fault units can be located through path query and a candidate set can be formed. Based on further aggregation and commonality analysis, main suspect objects can be accurately identified, fault causes can be revealed and timely warning can be performed. In this way, the traditional production management mode is transformed into active problem positioning and process optimization, and production efficiency is significantly improved and fault risk is reduced through intelligent methods.

Description

Technical Field

[0001] This application relates to the field of pressure sensor production management, and more specifically, to a visual management system and method for the production of pressure sensors for new energy vehicles. Background Technology

[0002] The widespread adoption of new energy vehicles is driving the continuous upgrading of automotive component manufacturing technology. Pressure sensors, as key components, play an indispensable role in the development of new energy vehicles. Pressure sensors are widely used in multiple core areas of new energy vehicles, such as power systems, braking systems, and suspension systems, and their quality and performance directly affect the safety and stability of the vehicle. However, due to the higher performance requirements of new energy vehicles, the manufacturing process of their internal pressure sensors has become increasingly complex. Traditional production management methods struggle to monitor key parameters in the production process in real time and cannot quickly locate potential fault sources when quality problems occur, resulting in a failure to effectively improve product yield and bottlenecks in production efficiency.

[0003] With the development of Industrial Internet technology, realizing the digitalization, visualization, and intelligent management of production processes through IoT technology has become a key trend in the upgrading of the manufacturing industry. For pressure sensor production lines, given the involvement of various equipment, parameters, product identification, and flow data, how to fully utilize this data to improve the transparency of the production process and achieve rapid identification of quality problems has become a pressing issue for the industry. Summary of the Invention

[0004] To address the aforementioned technical problems, this application is proposed. Embodiments of this application propose a visual management system and method for the production of pressure sensors for new energy vehicles. Through data collection, correlation, analysis, and visualization, it transforms the traditional production management model into proactive problem identification and process optimization. Through intelligent methods, it significantly improves production efficiency, reduces failure risks, and promotes the digitalization and high-quality development of new energy vehicle component production.

[0005] According to one aspect of this application, a method for visual management of pressure sensor production in new energy vehicles is provided, comprising: real-time collection of working parameters, equipment status data, product identification and circulation data, and test result data of key workstations through an IIoT gateway deployed on the pressure sensor production line; constructing pressure sensor production map data based on the working parameters, equipment status data, product identification and circulation data, and test result data; obtaining input quality alarms; querying the pressure sensor production map data for a candidate set of potential fault unit path data matching the quality alarms; aggregating and performing commonality analysis on the candidate set of potential fault unit path data to obtain the main suspected objects; and displaying the main suspected objects on a display screen and triggering a warning signal.

[0006] In one possible implementation, pressure sensor production graph data is constructed based on operating parameters, equipment status data, product identification and circulation data, and test result data. This includes: transmitting the operating parameters, equipment status data, product identification and circulation data, and test result data to a local data center and storing them in a time-series database, wherein the time-series database is indexed by product ID / batch ID, equipment ID, and parameter name; defining the node types and relationship types of the graph database, and updating the nodes and relationships in the graph database according to the operating parameters, equipment status data, product identification and circulation data, and test result data to obtain the pressure sensor production graph data.

[0007] In one possible implementation, the node type includes product unit, batch, process, equipment, process parameter, test item, and defect; the relationship type includes belonging to batch, passing through process, using equipment, generating parameters, performing tests, and associated defects.

[0008] In one possible implementation, querying the pressure sensor production map data for a candidate set of potential fault unit path data that matches the quality alarm includes: querying the pressure sensor production map data for a target node that matches the quality alarm; and traversing the pressure sensor production map data in reverse from the target node to obtain the candidate set of potential fault unit path data.

[0009] In one possible implementation, querying the pressure sensor production chart data for a target node that matches the quality alarm includes: extracting product unit, batch, test item, and defect from the quality alarm, and using the product unit in the quality alarm as a query node; and querying the pressure sensor production chart data for a target node that matches the query node.

[0010] In one possible implementation, traversing the pressure sensor production graph data backwards from the target node to obtain a candidate set of potential fault unit path data includes: traversing the pressure sensor production graph data backwards from the target node along the relationships between nodes to progressively identify nodes that match the batches, test items, and defects in the quality alarm to obtain the candidate set of potential fault unit path data.

[0011] In one possible implementation, the candidate set of potential fault unit path data is aggregated and subjected to commonality analysis to obtain the main suspect object, including: counting the number of common nodes and common relationships in the candidate set of potential fault unit path data to obtain statistical analysis results of common nodes and statistical analysis results of common relationships; taking the common node with the largest value in the statistical analysis results of common nodes as the main suspect object; and / or taking the common relationship with the largest value in the statistical analysis results of common relationships as the main suspect object.

[0012] In one possible implementation, the method of counting the number of shared nodes and shared relationships in the candidate set of the potential fault unit path data to obtain the statistical analysis results of shared nodes and shared relationships further includes: correcting the number of shared nodes and shared relationships in the candidate set of the potential fault unit path data to obtain the corrected number of shared nodes and shared relationships, and using the corrected number of shared nodes and shared relationships as the statistical analysis results of shared nodes and shared relationships.

[0013] In one possible implementation, correcting the number of shared nodes and shared relationships in the candidate set of potential fault unit path data to obtain a corrected number of shared nodes and shared relationships includes: dividing the statistical number of shared nodes and shared relationships by the total number of all nodes and all relationships to obtain node probability values ​​and relationship probability values; obtaining a binary discrete state function representing node activity and relationship activity based on the node probability values ​​and the relationship probability values; calculating a topology constraint value and a shared node / relationship density value based on the binary discrete state function representing node activity and relationship activity; calculating a number correction coefficient based on active coverage based on the topology constraint value and the shared node / relationship density value to obtain a first number correction coefficient and a second number correction coefficient; and correcting the number of shared nodes and shared relationships using the first number correction coefficient and the second number correction coefficient to obtain a corrected number of shared nodes and shared relationships.

[0014] According to another aspect of this application, a new energy vehicle pressure sensor production visualization management system is provided, used to execute the aforementioned new energy vehicle pressure sensor production visualization management method, comprising: a data acquisition module, used to collect in real time working parameters, equipment status data, product identification and circulation data, and test result data of key workstations through an IIoT gateway deployed on the pressure sensor production line; a production map data construction module, used to construct pressure sensor production map data based on working parameters, equipment status data, product identification and circulation data, and test result data; a quality alarm acquisition module, used to acquire input quality alarms; a potential fault unit query module, used to query the pressure sensor production map data for a candidate set of potential fault unit path data matching the quality alarm; a main suspect object determination module, used to aggregate and perform commonality analysis on the candidate set of potential fault unit path data to obtain the main suspect object; and an early warning prompt module, used to display the main suspect object on a display screen and trigger an early warning prompt signal.

[0015] Compared to existing technologies, the new energy vehicle pressure sensor production visualization management system and method provided in this application, by deploying IIoT gateways in the production line, achieves real-time collection of working parameters, equipment status, product identification, flow data, and test result data, comprehensively digitizing key information in the production process. Simultaneously, by constructing pressure sensor production graph data, production data is mapped into a graph database structure, thereby forming a flexible and efficient data relationship management and analysis mechanism. Utilizing the graph data structure, quality alarms can be responded to quickly, and potential faulty units can be located and candidate sets formed through path queries. Based on further aggregation and commonality analysis, the main suspects can be accurately identified, the cause of the fault can be revealed, and timely warnings can be issued. In this way, the traditional production management model is transformed into proactive problem localization and process optimization, significantly improving production efficiency and reducing fault risks through intelligent methods. Attached Figure Description

[0016] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1 The illustration shows a schematic flowchart of a method for visual management of the production of pressure sensors for new energy vehicles according to an embodiment of this application.

[0018] Figure 2 The illustration shows a schematic flowchart of step S2 in the new energy vehicle pressure sensor production visualization management method according to an embodiment of this application.

[0019] Figure 3 The illustration shows a schematic flowchart of step S4 in the new energy vehicle pressure sensor production visualization management method according to an embodiment of this application.

[0020] Figure 4 The illustration shows a schematic flowchart of step S41 in the new energy vehicle pressure sensor production visualization management method according to an embodiment of this application.

[0021] Figure 5 The illustration shows a schematic flowchart of step S5 in the new energy vehicle pressure sensor production visualization management method according to an embodiment of this application.

[0022] Figure 6 The figure shows a schematic block diagram of a new energy vehicle pressure sensor production visualization management system according to an embodiment of this application. Detailed Implementation

[0023] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0024] Figure 1 The illustration shows a schematic flowchart of a visual management method for the production of pressure sensors for new energy vehicles according to an embodiment of this application. Figure 1 As shown, this application provides a visualization management method for the production of pressure sensors for new energy vehicles, including: S1: Real-time collection of working parameters, equipment status data, product identification and circulation data, and test result data of key workstations through an IIoT gateway deployed on the pressure sensor production line; S2: Constructing pressure sensor production map data based on working parameters, equipment status data, product identification and circulation data, and test result data; S3: Obtaining input quality alarms; S4: Querying the pressure sensor production map data for a candidate set of potential fault unit path data that matches the quality alarms; S5: Aggregating and performing commonality analysis on the candidate set of potential fault unit path data to obtain the main suspected objects; S6: Displaying the main suspected objects on a display screen and triggering a warning signal.

[0025] Specifically, in step S1, the IIoT gateway deployed on the pressure sensor production line collects in real-time operating parameters, equipment status data, product identification and flow data, and test result data from key workstations. It should be understood that the production of pressure sensors for new energy vehicles involves numerous manufacturing and testing stages, with complex processes and high sensitivity to uncontrollable factors in the production process. To achieve refined quality management and efficient traceability, data transparency throughout the entire production process is essential. The fundamental prerequisite for achieving data transparency and visualization lies in the seamless, real-time, and automated collection of dynamic data from all key workstations in the production process. The Industrial Internet of Things (IIoT) gateway deployed on the production line is the core infrastructure supporting this work. This data includes not only the operating status and parameters of the production equipment itself, but also the unique identifier associated with each specific product and its flow process, as well as information on detection and testing results at each key stage.

[0026] In one embodiment, IIoT gateways deployed on the pressure sensor production line collect real-time operating parameters, equipment status data, product identification and flow data, and test result data for key workstations. This includes: firstly, a thorough analysis of the key workstations on the production line, specifically including chip mounting, soldering and wire bonding, potting and curing, calibration, final functional testing, and airtightness testing, etc. These processes determine the functional stability and final product quality. Each workstation is equipped with automated industrial control equipment and detection sensors, such as mounting machines, automatic soldering machines, curing ovens, and testing stations. Each device is natively equipped with or connected to various process parameter measurement and acquisition modules, such as temperature sensors, pressure sensors, flow transmitters, and current and voltage signal acquisition instruments. These modules output real-time equipment status data and operating process parameters. The equipment status data includes power on / off status, fault alarms, and self-diagnostic codes. The operating process parameters include heating temperature, pressure value, pressure holding time, and vibration amplitude. All of these data are transmitted externally through standard industrial interfaces, which include protocols such as RS485, Ethernet, CAN, Modbus RTU / TCP, and OPC-UA.

[0027] Furthermore, as a node aggregation device between physical devices and the upper-layer data platform, the IIoT gateway supports the conversion and compatibility of multiple industrial interconnection protocols. Each IIoT gateway, through configurable settings, can proactively poll real-time parameter data from production equipment and can also proactively push data change events from the devices to the gateway based on an industrial subscription / publishing mechanism. The types of data collected are not limited to data from physical sensors, but also include process and action records from equipment controllers (such as PLCs and embedded MCUs), such as start-up and shutdown times, operator IDs, alarm events, and equipment switching records. The data collection frequency is dynamically set according to equipment and process characteristics: for parameters with high dynamic fluctuations (such as pressure and temperature), the sampling accuracy can reach multiple times per second; for stable state data, flow records, and operation logs, collection is based on production cycle time and event-driven methods.

[0028] To collect product identification and flow data, industrial-grade barcode scanners, industrial cameras, or RFID readers are deployed at each workstation to ensure that each pressure sensor unit or its sub-component has a unique identification for tracking as it passes through all workstations. In one specific embodiment, a unique serial number (SN) or batch ID is assigned to each material upon arrival at the initial workstation. This serial number is then automatically collected and written using product barcodes, QR codes, or RFID tags. After each process step, the flow record is automatically uploaded to the IIoT gateway, including the timestamp of entry into the workstation, the time of departure from the workstation, the flow sequence number, and the process parameter template ID. This not only establishes a full lifecycle traceability chain for the product flow path but also enables layered collection and isolation of parallel production data across multiple workstations and batches.

[0029] For the automatic acquisition of test result data, each test station or automated quality control station is connected to a gateway. Through communication protocols with the test equipment, it not only collects the judgment data of the test results, such as pass / fail and NG reason codes, but also obtains all the raw test data directly related to the test. The raw test data includes electrical performance curves, pressure change curves, and sealing leakage values. Each test data packet is tagged with metadata information such as a unique product ID, related processes, acquisition time, and test equipment ID, enabling the data to be highly correlated and integrated with other production information in the backend.

[0030] After data collection at the network layer and edge, the IIoT gateway performs preliminary local processing on the raw data, such as multi-source format standardization, redundancy removal, sampling noise reduction, data packet loss compensation, and automatic error retransmission, to improve data quality and production security. Simultaneously, to ensure the reliability of the data reporting link, the IIoT gateway integrates a cache queue and a breakpoint resumption mechanism. Data can be temporarily stored locally during weak network conditions or network interruptions, and uploaded in batches when communication is restored, eliminating the risk of data loss caused by physical and systemic interruptions. After the above local preprocessing, the IIoT gateway pushes the collected data to the enterprise's local data center or cloud platform in real time at the millisecond level via industrial Ethernet, 5G networks, or dedicated metropolitan area IoT. The data is encapsulated and encrypted using standard industrial IoT protocols (such as MQTT, AMQP, CoAP, HTTPS, etc.) to ensure the real-time performance and security of enterprise production data.

[0031] At the data center, this collected, preprocessed, and multi-source integrated production data enters a dedicated data receiving and storage platform. The system distributes different data types into a time-series database and an object database according to structured rules. The time-series database carries all streaming data closely related to timestamps, such as equipment parameters, status change curves, and raw test waveforms. The data index uses a composite primary key of product ID, equipment ID, workstation ID, parameter type, and time interval to enable full-dimensional on-demand queries. The object database carries structured information such as flow records, process paths, and test judgments, spanning all process nodes within the production lifecycle.

[0032] While collecting data, the IIoT gateway also possesses local intelligent analysis capabilities. For example, it can perform edge-side rule-based judgments on fluctuations in operating parameters and device status, detect anomalies immediately, locally mark abnormal events, or proactively trigger upstream system alarms to quickly bring important events online. The data acquisition software uses configurable data adapters to map raw data fields from various acquisition devices to standardized interfaces, ensuring the uniformity and compatibility of data entry and subsequent retrieval.

[0033] In a specific embodiment, taking the pressure sensor welding station as an example, the equipment is connected to a temperature sensor and a precise current acquisition module. The IIoT gateway periodically (e.g., every 100ms) accesses the PLC to obtain real-time heating temperature and current curves, while simultaneously detecting the barcode scanner information at the station to identify the serial number and batch number of the sensor currently being processed. Each time the welding process is completed, the barcode data and process parameters are written to the database and immediately uploaded to the local central system. The system groups data from multiple parallel devices in the same process by index, ensuring that the production history information of any single pressure sensor is stored completely and accurately and can be quickly retrieved later. Similarly, the final testing station automatically detects airtightness. Through communication with the testing equipment's data bus, it synchronously uploads all original information such as test judgments, airtightness data curves, and anomalies. Each piece of data is precisely associated with a unique product ID and bound to metadata such as acquisition time, testing equipment, operator, and process template, providing accurate data support for subsequent problem localization and root cause analysis.

[0034] Specifically, in step S2, a pressure sensor production map is constructed based on operating parameters, equipment status data, product identification and flow data, and test result data. It should be understood that pressure sensors are core components in new energy vehicles with extremely high safety and reliability requirements, and every step in their production process can potentially affect the final product's performance and quality. In actual industrial scenarios, pressure sensor production lines often encompass highly automated process flows, complex equipment systems, and stringent testing standards. Relying solely on linear data acquisition and static storage cannot reveal the complex dynamic relationships between various stages of the production process, such as equipment parameters, process links, product batches, and testing procedures. Therefore, this application constructs a production map based on traceable data, thereby achieving full visibility of the production flow, efficient and convenient problem localization, and intelligent analysis of anomaly correlations.

[0035] In one embodiment, such as Figure 2 As shown, pressure sensor production graph data is constructed based on operating parameters, equipment status data, product identification and circulation data, and test result data, including: S21: transmitting operating parameters, equipment status data, product identification and circulation data, and test result data to a local data center and storing them in a time-series database, wherein the time-series database is indexed by product ID / batch ID, equipment ID, and parameter name; S22: defining the node types and relationship types of the graph database, and updating the nodes and relationships in the graph database according to the operating parameters, equipment status data, product identification and circulation data, and test result data to obtain the pressure sensor production graph data.

[0036] First, operating parameters, equipment status data, product identification and circulation data, and test results are transmitted to a local data center and stored in a time-series database, forming a full-cycle dynamic record for different times, different equipment, and different individual products or batches. This basic storage ensures that parameters at multiple time points and from multiple perspectives for any product unit or production process node have traceability, providing a solid data foundation for subsequent data modeling. More significantly, this allows for the definition of pressure sensor production graph data. Supported by a graph database, all important entities and their relationships are modeled using a node and edge structure. This operation transforms the data from a sequential log into a highly structured and interconnected network.

[0037] Next, the node types and relationship types of the graph database are defined, and the nodes and relationships in the graph database are updated according to working parameters, equipment status data, product identification and circulation data, and test result data to obtain the pressure sensor production graph data. In one embodiment, the node types include product unit, batch, process, equipment, process parameters, test items, and defects. It should be understood that the setting of node types achieves full coverage of the entire production process and all dimensions. Specifically, each pressure sensor is an independent product unit before leaving the factory, and all products produced in the same shift, the same batch, and with the same raw materials belong to one batch; each production line / production process link is defined as a process; each piece of equipment is equipment in actual production; at the same time, parameters that affect the process level and product performance (such as temperature, pressure, duration, flow rate, etc.) are defined as process parameters; the comprehensive inspection task for product quality is the test item; and all found non-conformity judgments and defect classification results are set as defects.

[0038] After establishing the aforementioned entity nodes, it is necessary to further define the types of relationships between them. These relationship types include batch, process, equipment used, parameter generation, test execution, and associated defect. Specifically, for example, product SN0001 on a production line belongs to batch B20240601. Each production and testing process it undergoes—such as surface mounting, soldering, potting, airtightness testing, and finished product marking—is recorded as a node within its lifecycle, establishing a process-through relationship between the product unit and the process. Each process is equipped with production equipment, such as soldering machines, mounting machines, and curing ovens, and this is recorded in the equipment-use relationship between the process and the equipment. Key process parameters collected during process execution (such as temperature, current, and pressure) are bound to the equipment or process through parameter generation relationships. A specific testing step (such as airtightness testing) is linked to the corresponding test item through test execution relationships. If the final product fails a test, such as detecting a leak, an associated defect relationship is created between the product unit and the defect node, and the non-conformance judgment and anomaly dimension are stored.

[0039] In a specific embodiment, taking a pressure sensor product unit SN10086 as an example, the data entry and graph modeling process of its entire production process can be summarized as follows: Upon product launch, it is bound to batch B20240601. After scanning the code, an edge is established between the product unit, the batch it belongs to, and the batch. When the product enters the surface mount process, the current device ID SPM-01 is detected, and the process and device node are recorded to form a process-device relationship. During surface mount, the sensor collects and transmits data in real time, including temperature, feed rate, and dwell time. All these parameters are aggregated and entered into the database via the IIoT gateway. Subsequently, regardless of the completion determination of each process, equipment malfunctions, process parameter anomalies, or the final test data and results, the product unit is the central node. Based on the data content, the relationships between various edges with other process, parameter, device, test item, and defect nodes are automatically mapped, supplemented, and improved.

[0040] Under this production diagram structure, the collection of each piece of production data is not merely an isolated value; it forms a networked connection throughout the entire product lifecycle, batch parent-child relationships, detailed process paths, equipment operating status, key process parameters, and even the final quality performance. For example, if a leak is found in product SN10086 during the final airtightness test, the product unit—process—equipment used—parameters generated—test executed—defect associated network structure can be followed to trace the product's status at each stage and the changes in equipment operating parameters, quickly converging to the possible root cause of the problem, achieving multi-dimensional diagnosis and attribution.

[0041] Specifically, in step S3, input quality alarms are acquired. It should be understood that on highly automated production lines, manual sampling and periodic analysis alone are often insufficient to promptly detect product defects and potential risks caused by complex processes and rapid mass production. By collecting and analyzing real-time data from each workstation, anomalies in the production process can be detected immediately, such as parameter exceedances, equipment malfunctions, or non-compliant test results. Simultaneously, these quality alarms are automatically integrated into the visual management system, thereby making subsequent data mining, correlation analysis, root cause analysis, and decision support more efficient and accurate.

[0042] In one embodiment, acquiring input quality alarms includes: In a pressure sensor production line, each key process and testing step carries a large amount of data reflecting product quality and equipment health. At testing stations, such as the final inspection stage, airtightness testing station, or electrical performance testing bench, automated testing instruments output pass / fail judgments, detailed values, anomaly codes, and specific defect categories for each product unit. An IIoT gateway or host computer system automatically collects all test data and its metadata, including the product's unique identifier, test time, test item, and device ID, and pushes it to the data platform in real time. Based on these highly timely data streams, the data center's backend system sets multi-level thresholds and rules, or applies a dynamic anomaly identification mechanism based on statistical algorithms and machine learning models, to automatically review and judge the collected data. When situations occur, such as a pressure sensor's airtightness test result being unqualified, leakage values ​​exceeding the allowable range, sudden parameter fluctuations during production, equipment reporting self-protection or abnormal shutdown, or multiple consecutive products exhibiting anomalies in the same batch and at the same workstation, a structured quality alarm can be automatically generated. The quality alarm includes information such as alarm ID, product unit, batch information, generation stage, abnormal parameters, test time, test item, defect type and severity level, and is pushed to the visual management system through an interface.

[0043] In other embodiments of this application, quality alarms can also be entered manually or by external quality monitoring agencies. If an operator discovers frequent anomalies in a batch of products on the monitoring interface, they can manually fill out an electronic form to submit a quality alarm. This manual or external data is then integrated into the production line big data system in a consistent and compatible manner, achieving unified management of quality alarms across all channels.

[0044] Specifically, in step S4, the candidate set of potential fault unit path data matching the quality alarm is queried from the pressure sensor production graph data. It should be understood that each quality alarm generated in the production process of pressure sensors for new energy vehicles actually corresponds to a specific product's non-conformity, failure, or potential fault event. However, these abnormal events are not isolated from a single data point or process node, but are often rooted in specific processes, links, or batches within the entire product manufacturing chain. These anomalies may originate from multiple factors such as raw material defects, parameter drift during processes, latent failures of key equipment, or insufficient verification after batch process adjustments. Therefore, only by organizing different types of real-time production data into a graph structure with correlations can a multi-dimensional, linked, and causally related systematic tracing of fault events be achieved.

[0045] In one embodiment, such as Figure 3As shown, the process of querying the pressure sensor production map data for a candidate set of potential fault unit path data that matches the quality alarm includes: S41: querying the pressure sensor production map data for a target node that matches the quality alarm; S42: traversing the pressure sensor production map data in reverse from the target node to obtain the candidate set of potential fault unit path data.

[0046] In one embodiment, such as Figure 4 As shown, the process of querying the pressure sensor production chart data and matching the target node with the quality alarm includes: S411: extracting product unit, batch, test item and defect from the quality alarm, and using the product unit in the quality alarm as the query node; S412: querying the pressure sensor production chart data and matching the query node with the target node.

[0047] Specifically, when a quality alarm is input, the product unit is first extracted from the quality alarm. Furthermore, the same alarm event also includes batch information, connecting product nodes to batch nodes through batch-related edges. Simultaneously, because the alarm involves inspection, non-conformance, or failure indicators, it also includes test item and defect type information, which can be synchronously used as location attributes for production graph nodes to enhance the relevance of node filtering and path aggregation during subsequent reverse traversal.

[0048] Next, the analysis process is initiated with the product unit in the pressure sensor production diagram as the target node. Based on the node relationships in the product lifecycle, the process is traversed upwards and backwards along multiple links, including processes, equipment usage, parameter generation, and test execution. Specifically, in one embodiment, the pressure sensor production diagram data is traversed backwards from the target node to obtain a candidate set of potential fault unit path data. This includes: traversing the pressure sensor production diagram data backwards from the target node along the relationships between nodes to progressively identify nodes that match the batch, test item, and defect in the quality alarm to obtain the candidate set of potential fault unit path data. This method, which uses the target node as a starting point and leverages the out-degree and in-degree relationships of nodes for multi-path, multi-angle backward traversal, comprehensively captures the process links, equipment parameters, and process distribution related to abnormal products, achieving multi-dimensional, link-based, and precise fault tracing, greatly improving the accuracy and efficiency of root cause investigation.

[0049] In a specific embodiment, taking the failure of a pressure sensor's airtightness test as an example, SN001122 is identified as the product unit in the pressure sensor production diagram. It is found that its batch edge points to Batch20240615, and its process edge sequentially points to welding, assembly, potting, and finally the airtightness test. Along this production path, each level of equipment can locate nodes such as welding machine AQ-05, potting equipment GF-03, and testing equipment TM-02. The parameter edges generated at each process stage can be tracked to the corresponding moment's parameters and values, such as welding temperature, welding current, equipment load, and motor speed. Simultaneously, the test execution edge can be associated with all test types performed, regardless of whether the result is pass or fail. Through reverse tracing and path traversal, all process procedures, equipment configurations, real-time process parameter statuses, execution operator codes, and related information of other products within the same batch related to this product are linked.

[0050] Furthermore, considering that the generation of candidate sets for potential fault unit path data cannot rely solely on data from a single instance, it is necessary to jointly consider similar anomalies under the same batch, process, equipment, or process parameters. In the graph data, combining the test items and defect type attributes of alarm nodes, by retrieving all other product unit nodes that share the same batch node with the target product unit, and then along the corresponding processes, equipment used, and parameters generated, we can check whether abnormal reports or parameter drifts occur in adjacent time periods, on the same equipment, or when key parameters exceed limits. Common candidate paths are defined as a set of links that may share common causes of the fault. For example, if it is found that in the same batch of products, products that have undergone the GF-03 potting process have multiple non-compliance items in the gas leakage test item, then the GF-03 potting process and production environment parameters (such as silicone temperature and curing time) under this path are included as high-priority paths in the candidate set. Similarly, if the testing equipment TM-02 frequently produces defective products over a period of time, the equipment node, its usage process, and the parameters at that time are all included in the candidate set for aggregation. In this way, data-driven intelligent investigation can avoid focusing only on a single product event and achieve rapid focus on potential systemic problems.

[0051] Specifically, in step S5, the candidate set of potential fault unit path data is aggregated and subjected to commonality analysis to obtain the main suspects. It should be understood that traditional methods for analyzing quality problems in the pressure sensor manufacturing process often rely on single-point anomaly analysis or manual tracing, which is insufficient to handle the cross-influence of data from multiple batches, workstations, and devices in large-scale, complex processes and high-concurrency production environments. Since anomalies in actual factory scenarios often have complex causes and dispersed manifestations, more scientific and intelligent source tracing analysis is needed for large and closely related datasets. Based on this, this application aggregates and performs commonality analysis on the candidate set of potential fault unit path data to utilize large-scale, structured production graph data to perform statistical induction and cross-analysis on the retrieved potential fault unit path candidate set. By discovering frequently co-occurring nodes and relationships, the true root cause of the fault is identified.

[0052] In one embodiment, such as Figure 5 As shown, the aggregation and commonality analysis of the candidate set of potential fault unit path data to obtain the main suspect objects includes: S51: counting the number of common nodes and common relationships in the candidate set of potential fault unit path data to obtain the statistical analysis results of common nodes and common relationships; S52: taking the common node with the largest value in the statistical analysis results of common nodes as the main suspect object; and / or, taking the common relationship with the largest value in the statistical analysis results of common relationships as the main suspect object.

[0053] Specifically, the candidate path set is based on quality alerts. It is obtained by reverse traversing the production graph data and filtering based on relevant conditions, capturing all node-relationship combinations corresponding to the alert. Each path represents a sequence of nodes and relationships, starting from the abnormal product unit node and traversing upstream of its complete production cycle, sequentially passing through all related processes, key equipment, process parameters, test items, and other entity nodes and process relationships. These paths collectively provide detailed data support and a structured description of all links that may lead to the product defect.

[0054] However, it is difficult to determine the systemic common causes of problems through a single path for a single product. Most anomalies in actual manufacturing are localized or sporadic phenomena, and isolated paths alone cannot reveal global risks. Only when multiple anomaly paths from different products, batches, and time periods highly overlap and converge at several key nodes or relationships can it indicate the existence of a process bottleneck, systemic failure, or that a certain process, equipment, or parameter has become the main cause of batch defects. Therefore, aggregation and commonality analysis mainly involve comprehensive statistics on the candidate path set, that is, performing union statistics, quantitative analysis, and intersection screening on the node and relationship sets of all paths. The goal is to identify the entity nodes or relationship types that appear most frequently in the large sample set. These high-quantity nodes and relationships are the common nodes and relationships, and their large-scale co-occurrence in the entire anomaly tracing chain usually represents the systemic root cause or high-risk area driving the spread of anomalies.

[0055] In a specific embodiment, all candidate paths extracted through reverse traversal are first stored in a structured manner. Each path consists of a series of node-relationship-node-relationship-node links. For example, starting from product SN0001, its traceability path can be specifically: Product unit node → Belongs to batch → Batch node → Processes passed through → Potting process node → Equipment used → GF-01 equipment node → Parameter generated → Temperature parameter node (abnormal) → Test performed → Air tightness test item node → Associated defect → Leakage defect node. As the number of abnormal products increases, multiple similar but different paths continuously converge at the system backend. The graph calculation engine accumulates the number of nodes in all paths by type (such as equipment, process, parameter, test item, etc.) and relationship type (processes passed through, equipment used, parameter generated, etc.). For each node / relationship, the total number of times it is included or traversed in all candidate paths is recorded, and these recurring nodes and relationships are defined as common nodes and common relationships. The number of each common node or common relationship, that is, the total number of times it is hit in the candidate paths, is called the statistical analysis result of that node or relationship.

[0056] When performing a reverse graph traversal from the target node based on the relationships between nodes, the selection of the critical path can be achieved by mining the global correlation of nodes / relationships in the pressure sensor production graph data. However, when counting the number of nodes and relationships on the critical path, it is also desirable to optimize the coverage density of common nodes / relationships in the overall graph structure, thereby improving the accuracy of aggregation and commonality analysis.

[0057] Based on this, in another embodiment, the method of counting the number of shared nodes and shared relationships in the candidate set of the potential fault unit path data to obtain the statistical analysis results of shared nodes and shared relationships further includes: correcting the number of shared nodes and shared relationships in the candidate set of the potential fault unit path data to obtain the corrected number of shared nodes and shared relationships, and using the corrected number of shared nodes and shared relationships as the statistical analysis results of shared nodes and shared relationships.

[0058] Specifically, the number of shared nodes and shared relationships in the candidate set of the potential fault unit path data is corrected to obtain the corrected number of shared nodes and shared relationships. This includes: firstly, dividing the statistical number of shared nodes and shared relationships by the total number of all nodes and all relationships to obtain node probability values ​​p1 and relationship probability values ​​p2, respectively; and based on the node probability values ​​and the relationship probability values, obtaining a binary discrete state function representing node activity and relationship activity, expressed as:

[0059] Φ=(p1,1-p1)

[0060] Ψ=(p2,1-p2)

[0061] Where Φ represents the binary discrete state function representing the node activity, and Ψ represents the binary discrete state function representing the relation activity.

[0062] Next, based on the binary discrete state function representing node activity and relation activity, the topological constraint value and the shared node / relation density value are calculated, expressed as:

[0063] α=||Φ-Ψ||2

[0064]

[0065] Where ||·||2 represents the L2 norm, α represents the topological constraint value, β1 represents the shared nodes, and β2 represents the relation density value. The topological constraint value α is used to implement set probability path selection constraints through Wasserstein-like distance calculation that reflects topological activity, while the shared node β1 / relation density value β2 is used to mark high-density traversal paths through high-density shared node / shared relation graph coverage.

[0066] Then, based on the topological constraint value and the shared node / relationship density value, the number correction coefficient based on active coverage is calculated to obtain the first number correction coefficient and the second number correction coefficient, expressed as:

[0067]

[0068] Wherein, ω1 represents the first number correction coefficient, and ω2 represents the second number correction coefficient.

[0069] Finally, the number of shared nodes and shared relationships is corrected using the first and second number correction coefficients to obtain the corrected number of shared nodes and shared relationships, expressed as:

[0070] n′1=ω1n1

[0071] n′2=ω2n2

[0072] Where n1 represents the number of shared nodes, n2 represents the number of shared relationships, n'1 represents the corrected number of shared nodes, and n'2 represents the corrected number of shared relationships.

[0073] In other words, the activity of nodes / relationships is used as a dynamic constraint entity in the traversal meta-path to ensure density coverage in the overall transition distribution of the graph structure. That is, the density aggregation degree of the overall probability space is quantified by probability measure to achieve dynamic adjustment of statistical characteristics based on commonality analysis, thereby improving the accuracy of the statistical analysis results of the common nodes and the statistical analysis results of the common relationships.

[0074] Furthermore, in this application, the nodes with the highest number of occurrences are identified as the primary suspects. In a specific embodiment, among the dozens of abnormal product sample paths obtained from the analysis, the GF-01 equipment node, the potting process node, and the abnormal curing temperature process parameter node appear repeatedly in most candidate paths, with their actual statistical counts being far higher than other nodes. Therefore, these nodes are considered the primary suspects most likely to be directly related to the current quality anomaly. Similarly, at the relationship analysis level, if the relationships involving the potting worker and the use of equipment GF-01 are hit with an extremely high number in most paths, or if several upstream and downstream relationships in the process flow appear with high density and repetition, then the corresponding relationship links should also be highly vigilant, and the nodes or chains they point to can be regarded as potential transmission channels for process runaway.

[0075] In other examples of this application, candidate paths are not only statistically analyzed in terms of quantity, but also integrated with multi-dimensional data such as time period, spatial distribution (e.g., different workstations, different production lines), and parameter change trends to improve the spatiotemporal accuracy of aggregated commonality analysis. For example, during the production of a pressure sensor batch B20240620, the traceability paths of 10 consecutive abnormal finished products were all hit at the potting process node and the GF-01 equipment node, and the number of abnormal curing temperature parameters collected also increased significantly. Based on such statistics, the number of hits at the potting process, GF-01 equipment, and corresponding parameter nodes will increase dramatically in a very short period of time. In this embodiment, in conjunction with actual workshop management strategies, judgment rules can be preset. For example, if the number of hits at a certain common node or common relationship accounts for more than 80% of the total number of candidate paths, it will be automatically identified as the main suspect. For more complex scenarios, the number of common relationships appearing under multiple link combinations can be recursively counted. For example, the link combination of potting process → GF-01 equipment → curing temperature appears frequently in all candidate paths and covers multiple products and batches, and is therefore identified as the main suspect.

[0076] Specifically, in step S6, the primary suspected object is displayed on the screen and a warning signal is triggered. It should be understood that in the complex process of producing pressure sensors for new energy vehicles, the amount of production data is enormous and changes extremely rapidly. Any drift in conditions, equipment failure, or process malfunction at any stage can lead to a large number of defective products. Therefore, timely, intuitive, and detailed feedback of fault analysis results to on-site operators, managers, and engineering teams is crucial to ensuring overall quality and collaborative efficiency. Based on this, in this application, the primary suspected object is displayed on the screen and a warning signal is triggered in a visual manner.

[0077] In one embodiment, after identifying the primary suspect, it is structurally encapsulated with relevant node attributes (such as equipment number, process name, batch number, parameter indicators, defect type, number of anomalies, and scope of impact) into a push object. This information is then integrated into a pre-defined monitoring and analysis module via the Manufacturing Execution System (MES), factory management information platform, or a customized data bus interface. On the display side, various presentation formats are employed, including large-screen real-time monitoring displays, regional management terminals, personal maintenance PCs, and even mobile apps or on-site HMI (Human-Machine Interface) terminals. Large screens and distributed terminals are deployed in the production monitoring center, process R&D office, and main entrances / exits of the factory workshop. The front end utilizes a visualization development framework or professional industrial visualization tools to construct a data display interface with multiple view dimensions, including entity topology diagrams, production flow paths, and anomaly statistics bar charts / heatmaps. Information on key suspects will be presented directly through dedicated sections, pop-ups, alert lists, or highlighted blocks. For example, the borders of suspected equipment will flash in the equipment hierarchy diagram, critical paths will be bolded in the process flow diagram, and warning signs will be displayed for the top-ranking abnormal nodes in the statistics table, ensuring that the information is prominent and the location is intuitive.

[0078] In summary, the new energy vehicle pressure sensor production visualization management method provided in this application, by deploying IIoT gateways in the production line, achieves real-time collection of working parameters, equipment status, product identification, flow data, and test result data, comprehensively digitizing key information in the production process. Simultaneously, by constructing pressure sensor production graph data, production data is mapped into a graph database structure, thereby forming a flexible and efficient data relationship management and analysis mechanism. Utilizing the graph data structure, quality alarms can be responded to quickly, and potential faulty units can be located and candidate sets formed through path queries. Based on further aggregation and commonality analysis, the main suspects can be accurately identified, the cause of the fault can be revealed, and timely warnings can be issued. In this way, the traditional production management model is transformed into proactive problem localization and process optimization, significantly improving production efficiency and reducing fault risks through intelligent methods.

[0079] This application also provides a new energy vehicle pressure sensor production visualization management system for executing the aforementioned new energy vehicle pressure sensor production visualization management method, such as... Figure 6As shown, the new energy vehicle pressure sensor production visualization management system 600 includes: a data acquisition module 601, used to collect in real time working parameters, equipment status data, product identification and circulation data, and test result data of key workstations through an IIoT gateway deployed on the pressure sensor production line; a production map data construction module 602, used to construct pressure sensor production map data based on working parameters, equipment status data, product identification and circulation data, and test result data; a quality alarm acquisition module 603, used to acquire input quality alarms; a potential fault unit query module 604, used to query the pressure sensor production map data for a candidate set of potential fault unit path data that matches the quality alarm; a main suspect object determination module 605, used to aggregate and perform commonality analysis on the candidate set of potential fault unit path data to obtain the main suspect object; and an early warning prompt module 606, used to display the main suspect object on a display screen and trigger an early warning prompt signal.

[0080] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0081] The flowcharts of the methods involved in this application are merely illustrative examples and are not intended to require or imply that connections, arrangements, or configurations must be made in the manner shown in the flowcharts. As those skilled in the art will recognize, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0082] It should also be noted that the steps in the method of this application can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions of this application.

[0083] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0084] The above description has been given for illustrative and descriptive purposes. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for visual management of pressure sensor production in new energy vehicles, characterized in that, include: The IIoT gateway deployed on the pressure sensor production line collects working parameters, equipment status data, product identification and circulation data, and test result data of key workstations in real time. Based on operating parameters, equipment status data, product identification and circulation data, and test result data, a pressure sensor production map data is constructed. Get input quality alerts; The candidate set of potential fault unit path data that matches the quality alarm is queried from the pressure sensor production graph data. The process of aggregating and performing commonality analysis on the candidate set of potential fault unit path data to obtain the main suspects includes: counting the number of shared nodes and shared relationships in the candidate set of potential fault unit path data; dividing the count of shared nodes and shared relationships by the total number of nodes and relationships to obtain node probability values ​​and relationship probability values; obtaining binary discrete state functions representing node activity and relationship activity based on the node probability values ​​and relationship probability values; calculating topological constraint values ​​and shared node / relationship density values ​​based on the binary discrete state functions representing node activity and relationship activity; calculating a number correction coefficient based on active coverage based on the topological constraint values ​​and the shared node / relationship density values ​​to obtain a first number correction coefficient and a second number correction coefficient; correcting the number of shared nodes and shared relationships using the first number correction coefficient and the second number correction coefficient to obtain a corrected number of shared nodes and shared relationships; selecting the shared node with the largest value in the statistical analysis results of the shared nodes as the main suspect; and / or selecting the shared relationship with the largest value in the statistical analysis results of the shared relationships as the main suspect. The main suspect is displayed on the screen, and a warning signal is triggered.

2. The method for visual management of new energy vehicle pressure sensor production according to claim 1, characterized in that, Based on operating parameters, equipment status data, product identification and circulation data, and test result data, a production map of pressure sensors is constructed, including: Working parameters, equipment status data, product identification and circulation data, and test results data are transmitted to the local data center and stored in a time series database, which is indexed by product ID / batch ID, equipment ID, and parameter name. Define the node types and relationship types of the graph database, and update the nodes and relationships in the graph database according to working parameters, equipment status data, product identification and circulation data, and test result data to obtain the pressure sensor production graph data.

3. The method for visual management of new energy vehicle pressure sensor production according to claim 2, characterized in that, The node types include product unit, batch, process, equipment, process parameter, test item, and defect; the relationship types include belonging to batch, passing through process, using equipment, generating parameters, performing tests, and associated defects.

4. The method for visual management of new energy vehicle pressure sensor production according to claim 1, characterized in that, The candidate set of potential fault unit path data that matches the quality alarm in the pressure sensor production graph data query includes: The target node that matches the quality alarm in the pressure sensor production graph data query; The pressure sensor production map data is traversed in reverse from the target node to obtain a candidate set of potential fault unit path data.

5. The method for visual management of new energy vehicle pressure sensor production according to claim 4, characterized in that, The target node that matches the quality alarm in the pressure sensor production data query includes: Extract product unit, batch, test item, and defect from the quality alarm, and use the product unit in the quality alarm as a query node; The target node that matches the query node in the pressure sensor production graph data query.

6. The method for visual management of new energy vehicle pressure sensor production according to claim 5, characterized in that, The candidate set of potential fault unit path data is obtained by traversing the pressure sensor production map data in reverse from the target node, including: The pressure sensor production graph data is traversed backwards from the target node along the relationships between nodes to progressively identify nodes that match the batches, test items, and defects in the quality alarms, thereby obtaining a candidate set of potential fault unit path data.

7. A visual management system for the production of pressure sensors for new energy vehicles, used to execute the visual management method for the production of pressure sensors for new energy vehicles as described in claim 1, characterized in that, include: The data acquisition module is used to collect working parameters, equipment status data, product identification and circulation data, and test result data of key workstations in real time through the IIoT gateway deployed on the pressure sensor production line; The production graph data construction module is used to construct pressure sensor production graph data based on working parameters, equipment status data, product identification and circulation data, and test result data. The quality alarm acquisition module is used to acquire input quality alarms; The potential fault unit query module is used to query the pressure sensor production graph data for a candidate set of potential fault unit path data that matches the quality alarm. The primary suspect identification module is used to aggregate and perform commonality analysis on the candidate set of potential fault unit path data to obtain the primary suspect, including: counting the number of shared nodes and shared relationships in the candidate set of potential fault unit path data; dividing the statistical number of shared nodes and shared relationships by the total number of all nodes and all relationships to obtain node probability values ​​and relationship probability values; obtaining binary discrete state functions representing node activity and relationship activity based on the node probability values ​​and relationship probability values; calculating topological constraint values ​​and shared node / relationship density values ​​based on the binary discrete state functions representing node activity and relationship activity; calculating a number correction coefficient based on active coverage based on the topological constraint values ​​and the shared node / relationship density values ​​to obtain a first number correction coefficient and a second number correction coefficient; correcting the number of shared nodes and shared relationships using the first number correction coefficient and the second number correction coefficient to obtain a corrected number of shared nodes and shared relationships; selecting the shared node with the largest value in the statistical analysis results of the shared nodes as the primary suspect; and / or selecting the shared relationship with the largest value in the statistical analysis results of the shared relationships as the primary suspect. The early warning module is used to display the main suspected object on the display screen and trigger an early warning signal.

Images