Control panel delivery intelligent quality inspection and tracing system

Abstract

A control panel delivery intelligent quality inspection and tracing system solves the problems that in the prior art, an automatic detection process is prone to being interrupted, data is dispersed and difficult to associate and trace, and dynamic risk early warning is lacked, and comprises an intelligent quality inspection subsystem and a delivery tracing subsystem which cooperate with each other. The quality inspection subsystem ensures safe and accurate detection through magnetic coupling isolation and golden finger connection, guarantees no loss of network disconnection data and network bandwidth optimization by using SQLite caching and a timed uploading mechanism, improves the maintenance and remodeling efficiency by adopting a physical separation detachable needle bed, and sets a fault simulation unit to comprehensively verify the reliability of a control panel. The traceability subsystem is in butt joint with cloud data through a Flask server, constructs an associated database to realize a full-process data link, supports macroscopic to microscopic data analysis by using a visual engine, performs anomaly identification and risk prediction in combination with machine learning, and generates an early warning report; according to the system, the delivery quality control level, the production efficiency and the risk prevention capability of the control panel are improved.

Description

Technical Field

[0001] This article belongs to the technical field of industrial automation quality control systems, specifically involving an intelligent quality inspection and traceability system for control boards leaving the factory. Background Technology

[0002] In the field of high-end electronic equipment manufacturing, the precision and reliability of printed circuit board control boards directly determine the performance of the entire machine. With the advancement of Industry 4.0, automated quality control systems have become a core component of high-precision PCB manufacturing, and their importance is mainly reflected in the following three aspects:

[0003] Quality Assurance Requirements: As the nerve center of electronic equipment, even micron-level deviations in the PCB control board can cause systemic failures. Real-time, closed-loop automated testing is necessary to ensure a product yield of ≥99.95%.

[0004] Production efficiency constraints: Traditional manual inspection has a speed of less than 5 pieces / minute and a miss rate as high as 8%, which cannot match the high-speed cycle of modern SMT production lines of ≥30 pieces / minute.

[0005] Process optimization depends on: quality control data is a key input for process iteration, and full-process data correlation needs to be achieved to support continuous improvement.

[0006] Despite the widespread application of Industrial Internet of Things (IIoT) technology, existing automated quality control systems still face three major technical bottlenecks:

[0007] 1. The automated process is broken. The current system relies on preset rules to determine anomalies and cannot dynamically respond to new fault modes. In complex scenarios, the rate of manual re-inspection exceeds 40%, which leads to an average detection cycle of 15 minutes per batch and an error rate of 5.2%.

[0008] 2. Data silos are severe; test data is scattered across local databases, log files, and the MES system, lacking a unified spatiotemporal labeling system.

[0009] The test results are not dynamically linked to the device ID, operator, or environmental parameters.

[0010] The timing data is not aligned with the production line cycle time;

[0011] Cross-process traceability relies on manually linking Excel spreadsheets, and fault location takes an average of ≥48 hours.

[0012] 3. The risk warning mechanism is lacking. The existing system only triggers alarms based on static thresholds and has no ability to predict hidden quality degradation. Historical data shows that 42% of batch failures have shown a trend of abnormality for 36 consecutive hours before they occur, but the existing technology cannot identify such signals.

[0013] To overcome the aforementioned bottlenecks, the high-precision PCB manufacturing industry urgently needs the following technological upgrades:

[0014] Full-process unmanned inspection: Develop an adaptive fault simulation engine to achieve dynamic learning and judgment of unknown defect types;

[0015] End-to-end data traceability: Construct a distributed storage architecture with spatiotemporal correlation to support millisecond-level correlation retrieval of test data, process parameters, and equipment status;

[0016] Dynamic risk prediction: Integrating machine learning algorithms to analyze implicit patterns in historical data and establishing predictive maintenance models. Summary of the Invention

[0017] This paper presents an intelligent quality inspection and traceability system for control boards, aiming to solve three core problems in existing technologies: broken automated testing processes, data silos, and lack of risk warning. Through the collaborative architecture of QT client system, Flask server system and QTC hardware testing system, the system realizes closed-loop management of the entire chain from signal acquisition to decision optimization.

[0018] Intelligent quality inspection system:

[0019] A smart quality inspection system for control boards before shipment, comprising the following core modules:

[0020] QTC Hardware Testing Module: Supports flexible configuration with multiple small boards, integrating an 8-channel analog voltage sampling circuit, an 8-channel digital potentiometer, dual RS232 interfaces, and magnetic coupling isolation devices. The primary side of the magnetic coupling isolation device connects to the sampling circuit, and the secondary side connects to the communication circuit, effectively isolating the high and low voltage sides, significantly improving the electrical safety and anti-interference capability of the system. This module connects to the control board under test through a gold finger connector, providing stable and reliable electrical contact and ensuring the accuracy of the test signal.

[0021] QT host computer module: running within an industrial computer, equipped with an SQLite database, CAN bus interface, and MQTT protocol communication interface. The SQLite database is responsible for local storage of test data, ensuring that critical data is not lost when the network is unstable or interrupted. The CAN bus interface is used to receive the detection voltage value, PWM duty cycle, and amplitude information reported by the QTC hardware detection module, realizing efficient and reliable data interaction with the hardware module. The MQTT protocol communication interface uploads data to the OneNet IoT cloud platform, facilitating remote monitoring and centralized data management.

[0022] This module also includes:

[0023] Timer unit: Triggers timed batch upload and incremental upload mechanisms, effectively optimizing network bandwidth usage and improving upload efficiency;

[0024] SQLite local cache database: temporarily stores data when the network is interrupted, and automatically resumes transmission after the network is restored, ensuring the integrity and continuity of test data;

[0025] Probe control module: Physically separated from the QTC hardware detection module, it receives instructions from the QT host computer through an electrical signal interface to drive the detachable needle bed adapter interface. This physical separation design simplifies the system structure and facilitates maintenance. The detachable needle bed design allows the system to quickly adapt to different models of control boards, greatly improving the efficiency of production line changeover. The physical specifications of its needle bed adapter interface are replaceable to adapt to the test point arrays of different models of control boards, greatly enhancing the system's versatility and production line adaptability, and significantly reducing equipment investment costs when producing multiple varieties.

[0026] Fault simulation unit: By applying controllable voltage and current to the control board under test through the QTC hardware detection module, it simulates faults such as temperature sensor and interlock, comprehensively verifies the response and protection mechanism of the control board under various fault conditions, and effectively improves the reliability and stability of the products leaving the factory.

[0027] Factory traceability system:

[0028] A control board manufacturing intelligent traceability system, comprising the following components:

[0029] Flask server module: Receives data uploaded to OneNet by the QT host computer module through a RESTful API interface, enabling efficient and standardized data exchange between the cloud and backend services;

[0030] Database storage unit: Associates storage device ID, operator information, timestamp, production line test data and shipment verification information to build a complete product lifecycle data chain, laying a solid foundation for accurate traceability;

[0031] The visualization engine unit generates real-time dashboards, historical data charts, and maintenance record interfaces, and supports queries by equipment, time, test point, and other criteria, providing intuitive, multi-dimensional data views that significantly improve the efficiency of quality status monitoring and problem analysis. This unit can switch between batch pass rate statistical charts and individual board maintenance record details based on user commands, allowing users to quickly focus from macro-quality trends to micro-level individual board issues, enabling rapid identification of the root cause of problems.

[0032] Machine learning early warning unit: Based on historical data, it dynamically identifies abnormal patterns, performs regression analysis and time series prediction, and outputs quality problem early warnings, effectively reducing batch quality risks and improving the overall product quality level;

[0033] Early warning report generation unit: Generates reports in PDF or Excel format and sends early warning notifications to designated personnel based on the SMTP email protocol, ensuring that quality problem early warning information can reach relevant responsible persons in a timely, accurate and multi-form manner, and accelerating the problem response and handling closed loop.

[0034] Beneficial effects:

[0035] The intelligent quality inspection and traceability system for control boards proposed in this paper effectively solves three major problems: automated detection failure, data silos, and lack of risk warnings. In the quality inspection stage, the magnetic coupling isolation and gold fingers of the QTC module ensure safety and signal accuracy, the SQLite cache of the QT host computer ensures that data is not lost when the network is down, the CAN and MQTT protocols enable efficient data interaction and uploading, the timer optimizes bandwidth, and the physically separated detachable needle bed greatly simplifies maintenance, improves changeover efficiency and multi-board type adaptation, and reduces costs. The fault simulation unit fully verifies the reliability of the control board. In the traceability stage, the Flask server efficiently connects to the cloud, the database builds a complete product data chain, the visualization engine supports rapid focusing from macro trends to micro problems, the machine learning unit dynamically identifies anomalies and predicts risks to prevent batch problems, and the early warning report ensures that information reaches the responsible person in a timely manner, accelerating the response loop. The system comprehensively improves quality control, production efficiency and risk prevention capabilities.

[0036] 1. Full-process test record traceability

[0037] Spatiotemporal correlation data chain:

[0038] The database storage unit uses a unique device ID as the primary key and dynamically binds operators, precise timestamps, production line test data (such as voltage values ​​and PWM duty cycles), and shipment verification information to build a full lifecycle data chain with millisecond-level precision.

[0039] Resumable downloads ensure integrity:

[0040] The SQLite local caching and scheduled upload mechanism of the QT host computer ensures zero data loss during network outages and automatically resumes uploading to the cloud after the network is restored, achieving 100% traceability of test records.

[0041] Precise fault location:

[0042] The visualization engine supports multi-dimensional queries by device ID, time range, test point, etc. It can locate the historical test data and maintenance records of abnormal boards within 30 seconds, improving fault location efficiency by more than 99% compared to traditional manual tracing (≥48 hours).

[0043] 2. Product lifecycle traceability

[0044] End-to-end data connectivity:

[0045] From needle bed testing (probe control module) and fault simulation (QTC module) to shipment verification (Flask server), data from all stages are linked through unified spatiotemporal tags to form an unalterable product traceability file.

[0046] Seamless macro-micro penetration:

[0047] The visualization engine allows for one-click switching between batch pass rate statistical charts and single-board details (such as...). Figure 7 Historical test data Figure 10 (Maintenance records) enable second-level focusing from batch trends to the root cause of single-board problems.

[0048] Dynamic risk source tracing and prediction:

[0049] The machine learning early warning unit performs regression analysis based on historical data (such as continuous 36-hour trend anomalies) to identify hidden quality degradation in advance, link and trace the IDs of devices in the same batch, and prevent 42% of potential batch failures.

[0050] 3. Quality control and efficiency improvement

[0051] Quality inspection process:

[0052] Magnetic coupling isolation and gold finger connection ensure zero signal interference, with detection accuracy down to the microvolt level;

[0053] The detachable needle bed (probe control module) allows for changeover within 10 minutes, is compatible with test point arrays of multiple control boards, and increases equipment reuse rate by 90%.

[0054] The fault simulation unit covers more than 20 fault scenarios, including temperature sensors and interlocks, to fully verify the reliability of the control board.

[0055] Traceability process:

[0056] Flask server-side enables cloud data connection in seconds via RESTful API;

[0057] The early warning report generation unit automatically pushes PDF / Excel reports to the responsible person, reducing the response time for quality issues to 5 minutes.

[0058] 4. Economic benefits

[0059] The automated detection rate reaches ≥30 pieces / minute, and the false positive rate is reduced to 0.1%;

[0060] Data silos are eliminated, reducing manual Excel workflows by 80%;

[0061] Predictive maintenance reduces batch quality incident losses by up to 60%. Attached Figure Description

[0062] Figure 1 This is a schematic diagram of the host computer interface of a QT system for a control board intelligent quality inspection and traceability system.

[0063] Figure 2 This is a schematic diagram of local file data in a QT system for a control board intelligent quality inspection and traceability system.

[0064] Figure 3 This is a first schematic diagram of the QTC system process of a control board intelligent quality inspection and traceability system for factory shipment.

[0065] Figure 4 This is a second schematic diagram of the QTC system flow of a control board intelligent quality inspection and traceability system;

[0066] Figure 5 This is a schematic diagram of the data instrument interface of a Flask system, which is a control board intelligent quality inspection and traceability system for factory shipment.

[0067] Figure 6 This is a schematic diagram of the real-time test data interface of a Flask system, which is a control board intelligent quality inspection and traceability system for factory shipment.

[0068] Figure 7 This is a schematic diagram of the historical test data interface of a Flask system, which is a control board intelligent quality inspection and traceability system for factory shipment.

[0069] Figure 8 This is a schematic diagram of the Flask system's shipment verification interface, which is a control board intelligent quality inspection and traceability system for factory shipment.

[0070] Figure 9 This is a schematic diagram of the Flask system's shipment record management interface, which is a control board intelligent quality inspection and traceability system for factory shipments.

[0071] Figure 10 This is a schematic diagram of the Flask system maintenance record management interface, which is a control board intelligent quality inspection and traceability system for factory shipment.

[0072] Figure 11 This is a schematic diagram of the Flask system file upload interface for a control board intelligent quality inspection and traceability system. Detailed Implementation

[0073] To enhance understanding of the present invention, the invention will be further described in detail below with reference to embodiments and accompanying drawings. These embodiments are only for explaining the invention and do not constitute a limitation on the scope of protection of the invention.

[0074] like Figure 1 , 2 As shown in 3, 4, 5, 6, 7, 8, 9, 10, 11

[0075] Intelligent quality inspection system:

[0076] A smart quality inspection system for control boards before shipment, comprising the following core modules:

[0077] QTC Hardware Testing Module: Supports flexible configuration of multiple small boards, integrates 8-channel analog voltage sampling circuit, 8-channel digital potentiometer, dual RS232 interface and magnetic coupling isolation device. The primary side of the magnetic coupling isolation device is connected to the sampling circuit and the secondary side is connected to the communication circuit. This module is connected to the control board under test through gold finger connector.

[0078] QT host computer module: running in an industrial computer, equipped with an SQLite database, a CAN bus interface and an MQTT protocol communication interface. The SQLite database is responsible for local storage of test data, the CAN bus interface is used to receive the detection voltage value, PWM duty cycle and amplitude information reported by the QTC hardware detection module, and the MQTT protocol communication interface uploads data to the OneNet IoT cloud platform.

[0079] This module also includes:

[0080] Timer unit: Triggers timed batch upload and incremental upload mechanisms;

[0081] SQLite local cache database: temporarily stores data when the network is interrupted and automatically resumes transmission when the network is restored;

[0082] Probe control module: Physically separate from the QTC hardware detection module, it receives instructions from the QT host computer through an electrical signal interface to drive the detachable needle bed adapter interface. The physical specifications of the needle bed adapter interface can be replaced to adapt to the test point array of different control boards.

[0083] Fault simulation unit: The QTC hardware detection module applies controllable voltage and current to the control board under test to simulate faults such as temperature sensor and interlock.

[0084] Factory traceability system:

[0085] A control board manufacturing intelligent traceability system, comprising the following components:

[0086] Flask server module: Receives data uploaded to OneNet by the QT host computer module via a RESTful API interface;

[0087] Database storage unit: associated storage device ID, operator information, timestamp, production line test data, and shipment verification information;

[0088] Visualization Engine Unit: Generates real-time dashboards, historical data charts, and maintenance record interfaces, and supports queries by equipment, time, test point, and other criteria. This unit can switch between batch pass rate statistics charts and single-board maintenance record details based on user commands.

[0089] Machine learning early warning unit: dynamically identifies abnormal patterns based on historical data, performs regression analysis and time series prediction, and outputs early warnings of quality problems;

[0090] Warning report generation unit: Generates reports in PDF or Excel format and sends warning notifications to designated personnel based on the SMTP email protocol.

[0091] Implementation Example

[0092] 1. Preparation and Testing:

[0093] Replace the needle bed: Depending on the model of the control board under test, replace the detachable needle bed adapter interface on the probe control module.

[0094] Wiring: Connect the bed of needles to the probe control module, connect the QTC hardware detection module to the control board using the gold finger connector, and then connect the probe module and QTC module to the industrial computer running the QT host computer software.

[0095] Start-up Test: The operator inputs device and operator information into the QT software to start the test. The QTC module automatically acquires control board voltage, PWM, and other signals, and transmits the data back to the QT software via the CAN bus. If necessary, the QTC module can simulate fault testing of the control board's response.

[0096] 2. Data processing and uploading:

[0097] Local storage: The QT software stores all received test data into a local SQLite database in real time, ensuring that data is not lost when the network is disconnected.

[0098] Optimized upload: With a timer-triggered mechanism, the QT software uploads local data in batches or incrementally to the OneNet cloud platform via the MQTT protocol, optimizing bandwidth. Data lost due to network outages will automatically resume transmission after the network is restored.

[0099] 3. Cloud-based traceability and monitoring:

[0100] Cloud-based reception and storage: The Flask server receives OneNet data via API, stores it in a cloud database, and associates it with the entire lifecycle information of the storage device to form a complete data chain.

[0101] Visual analytics: Quality personnel can view real-time dashboards, historical charts, and detailed records through the visualization engine, and can quickly switch between macro statistics and individual panel details.

[0102] 4. Intelligent early warning:

[0103] Machine learning analysis: The system dynamically identifies abnormal patterns based on historical data and performs predictive analysis.

[0104] Risk warning: Automatically generate warnings when potential quality risks are detected.

[0105] Reports and Notifications: Automatically generate PDF / Excel alert reports and send them to relevant personnel via email in a timely manner, accelerating the problem response loop.

[0106] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A factory-designed intelligent quality inspection system for control boards, characterized in that, include: The system comprises a QTC hardware detection module, a QT host computer module, a probe control module, and a fault simulation unit. The QTC hardware detection module supports flexible configuration of multiple small boards, including: an 8-channel analog voltage sampling circuit, an 8-channel digital potentiometer, dual RS232 interfaces, and magnetically coupled isolation devices; The magnetically coupled isolation device is connected to the sampling circuit on the primary side and to the communication circuit on the secondary side. The QTC hardware testing module is connected to the control board under test via a gold finger connector; The QT host computer module runs within an industrial computer and includes: an SQLite database, a CAN bus interface, and an MQTT protocol communication interface; The SQLite database is used to store test data locally; The CAN bus interface receives the detection voltage value, PWM duty cycle and amplitude information reported by the QTC hardware detection module; The MQTT protocol communication interface is used to upload data to the OneNet IoT cloud platform; The probe control module is physically separated from the QTC hardware detection module. It receives QT commands through an electrical signal interface and is controlled and driven by the QT host computer commands to drive the detachable needle bed adapter interface. The fault simulation unit applies controllable voltage and current to the control board under test through the QTC hardware detection module to simulate temperature sensor and interlock faults.

2. The intelligent quality inspection system for control boards as described in claim 1, characterized in that: The QT host computer module includes: a timer unit and an SQLite local cache database; The timer unit triggers the timed batch upload and incremental upload mechanisms; The SQLite local cache database temporarily stores data when the network is interrupted.

3. The intelligent quality inspection system for control boards as described in claim 1, characterized in that: The physical specifications of the probe control module's needle bed adapter interface are replaceable, allowing it to adapt to test point arrays of different control boards.

4. A control board factory traceability system, characterized in that, include: Flask server module, database storage unit, visualization engine unit, and machine learning early warning unit; The Flask server module receives data uploaded to OneNet by the QT host computer module via a RESTful API interface; The database storage unit is associated with storage device ID, operator information, timestamp, production line test data, and shipment verification information; The visualization engine unit generates real-time dashboards, historical data charts, and maintenance record interfaces, and supports query functions based on equipment, time, and test point conditions. The machine learning early warning unit dynamically identifies abnormal patterns based on historical data, performs regression analysis and time series prediction, and outputs early warnings of quality problems.

5. A control board factory traceability system according to claim 4, characterized in that: The visualization engine unit switches between batch pass rate statistics charts and single-board repair record details according to user instructions.

6. A control board factory traceability system according to claim 4, characterized in that: The traceability system also includes an early warning report generation unit, which generates PDF or Excel format reports and sends early warning notifications to designated personnel based on the SMTP email protocol.

Images