High-precision multi-interface collaborative FTU dynamic function test system and method

By introducing acrylic transparent protective covers, quick-change carrier boards, positioning components and various sensors into the FTU test system, combined with PLC module control, high-precision multi-interface collaborative testing is achieved, solving the problems of insufficient positioning accuracy, poor interface test coordination, and low safety protection reliability, improving the level of automation and intelligence, and meeting the quality control needs of smart grid equipment.

CN120750805APending Publication Date: 2025-10-03ZHUHAI GLEX AUTOMATION EQUIP CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202511136334.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-08-14
Publication Date
2025-10-03

AI Technical Summary

Technical Problem

The existing FTU test system has problems such as insufficient positioning accuracy, poor interface test coordination, low safety protection reliability, and limited automation and intelligence levels, and cannot meet the quality control needs of large-scale production of smart grid equipment.

Method used

The protective unit consists of a transparent acrylic protective cover and a sheet metal box, combined with a quickly replaceable carrier plate, access structure and positioning components, an integrated pressure plate structure, an endoscope and a test module, and is equipped with a safety grating, an anti-pinch device and a pre-feeding sensing mechanism. The test process is controlled by a PLC module, and a variety of sensors and servo motors are used to achieve dynamic adjustment and data fusion analysis.

Benefits of technology

It improves the test accuracy and reliability, test efficiency and automation level, ensures test accuracy and safety, achieves a significant improvement in the reliability and automation level of interface testing, generates a comprehensive scoring report, provides a significant improvement in the automation level of interface testing, generates a comprehensive scoring report for interface testing, ensures a significant improvement in the automation level of interface testing, generates a significant improvement in the automation level of interface testing, generates a significant improvement in the automation level of interface testing, generates a significant improvement in the automation level of interface testing, generates a significant improvement in the automation level of interface testing, and generates a significant improvement in the automation level of interface testing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120750805A_ABST
    Figure CN120750805A_ABST
Patent Text Reader

Abstract

The invention discloses a high-precision multi-interface collaborative FTU dynamic function test system and method. The system comprises a protection unit, a test bearing unit, a core test unit and a control system, the protection unit is composed of an acrylic transparent protection cover and a metal plate box body, and a test space is formed. The test bearing unit comprises a quickly replaceable carrier plate, an access structure and a positioning assembly, and the positioning assembly comprises a positioning pin and a guide block; the core test unit comprises a pressing plate structure, an endoscope and a test module; the test module is integrated with a key, a short circuit bolt, a network port, a DB9 and an aviation plug test module; the safety detection unit comprises a safety grating, an anti-pinch device and a pre-feeding induction mechanism, and the safety grating is arranged at the inlet of the test space; and the control system controls starting, execution and result uploading of the test process. The method is used for solving the technical problems of insufficient positioning precision, poor interface test collaboration, low safety protection reliability and limited automation and intelligence level in the existing FTU test.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of power system automation, and in particular to a high-precision multi-interface coordinated FTU dynamic function testing system and method. Background Art

[0002] In the field of power system automation, feeder terminal units (FTUs) are key equipment for monitoring and controlling distribution networks. Their functional reliability directly impacts grid stability. With the advancement of smart grid development, FTUs are integrating an increasing variety of interfaces (such as buttons, network ports, DB9 connectors, and aviation plugs), placing higher demands on test system compatibility, test accuracy, and automation. However, existing FTU functional testing technology still has the following significant limitations: 1. Design flaws in the test system structure Traditional test equipment often utilizes open frames or simple protective covers, lacking an enclosed test space. These devices are susceptible to external electromagnetic interference (EMI) and dust, leading to distortion of high-frequency signal test data (e.g., excessive jitter in network port eye diagrams). Test carriers are often fixed fixtures, making it difficult to adapt to the rapid changeover requirements of different FTU models. Replacing carriers can take over 30 minutes, hindering production line testing efficiency. Positioning components typically utilize a rigid pin-and-hole fit, which is unable to compensate for product manufacturing tolerances (±0.1mm) and cumulative installation errors. This results in a docking deviation rate exceeding 5% between the test module and the interface, posing a risk of mechanical collision.

[0003] 2. Insufficient performance of the core test module Keystroke testing often relies on a fixed-stroke cylinder drive, making it impossible to dynamically adjust the trigger pressure based on the elastic modulus of different keystroke materials, such as silicone and metal domes, leading to overvoltage damage or poor contact. Interface test modules (such as network ports and DB9) lack an integrated coordination mechanism. Mechanical insertion and removal force and electrical performance testing are performed separately, failing to capture dynamic parameter correlations during the contact process (such as the impact of contact resistance jumps on signal delays). This results in a high rate of missed detection of latent faults. Short-circuit pin testing utilizes an electromagnetic push rod drive with an unadjustable feed speed. This results in large fluctuations in contact resistance, leading to distortion of the short-circuit current waveform and affecting the accuracy of overcurrent protection algorithm verification.

[0004] 3. Low level of safety testing and automation Safety protection often relies on a single photoelectric sensor, which is susceptible to interference from lighting fluctuations, leading to false triggering and shutdown. The lack of a self-calibration mechanism for the optical path makes troubleshooting time-consuming. The testing process is also poorly automated, requiring manual intervention for key steps like lid status confirmation and multi-rate protocol switching, making it impossible to meet mass production requirements.

[0005] 4. Lack of standardization of test methods Existing testing methods inadequately simulate dynamic operating conditions. For example, DB9 interface load testing uses a fixed current value, which cannot replicate the dynamic current fluctuations of 0-500mA in real applications, resulting in overcurrent protection response time testing errors exceeding 20%. The lack of simultaneous monitoring of contact pressure, temperature, and high-frequency interference coupling during the aviation plug insertion process makes it impossible to assess interface reliability in complex electromagnetic environments. Test results rely on manual comparison and lack a quantitative analysis algorithm based on dynamic response curves, resulting in high subjectivity and poor consistency.

[0006] In response to the above problems, there is an urgent need to develop an FTU functional testing system and method with high integration, excellent test accuracy, strong automation and safety to meet the quality control needs of large-scale production of smart grid equipment. Summary of the Invention

[0007] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide a high-precision multi-interface collaborative FTU dynamic function testing system and method, which is used to solve the technical problems existing in the existing FTU testing, such as insufficient positioning accuracy, poor interface testing coordination, low safety protection reliability, and limited automation and intelligence levels.

[0008] In order to solve the above problems, the technical solutions adopted by the present invention are as follows: A high-precision multi-interface collaborative FTU dynamic function test system, including: Protection unit: It is composed of an acrylic transparent protective cover and a sheet metal box. The acrylic transparent protective cover covers the top of the sheet metal box to form a test space. Test carrier unit: includes a quickly replaceable carrier plate, an access structure, and a positioning assembly. The access structure is driven by a rodless cylinder to achieve the in and out movement of the carrier plate on the slide rail. The positioning assembly includes a positioning pin and a guide block. Core test unit: includes a pressure plate structure, an endoscope, and a test module. The pressure plate structure is driven by a three-axis cylinder to achieve lifting and pressing actions. The endoscope and test module are located below the pressure plate structure. The test module integrates a button, short-circuit pin, network port, DB9, and aviation plug test module. The button test module is driven by a needle-type cylinder. Safety detection unit: including safety grating, anti-pinch device and pre-feeding sensing mechanism. The safety grating is set at the entrance of the test space; Control system: controls the start-up, execution and result uploading of the test process, and is equipped with PLC modules and database storage functions.

[0009] A high-precision multi-interface collaborative FTU dynamic function testing method includes the following steps: Place the FTU product to be tested on the carrier, manually open the product lid, and trigger the pre-feed sensing mechanism to use a dual-wavelength collaborative detection mechanism for non-contact preliminary judgment, and then perform a secondary confirmation through physical contact to confirm that the lid is in the open state; Start the safety light barrier installed at the entrance of the test space, complete the optical path calibration, signal threshold configuration and linkage equipment communication test in sequence, and enter the detection state; The carrier plate is driven to move along the slide rail by a rodless cylinder. Positioning is performed by a positioning pin with a dual-stage floating structure integrated with a micro pressure sensor, and the positioning contact pressure is monitored in real time. Positioning deviations are automatically compensated by the fine-tuning screw of the guide block, and the carrier plate enters the test position. The three-axis cylinder drives the pressure plate structure downward, so that the buttons, short-circuit pins, network ports, DB9 and aviation plug test modules integrated in the pressure plate structure contact the interfaces of the FTU product to be tested in turn to perform performance testing; During the test, the endoscope collects test image data in real time, and the control system synchronously uploads the test data and image data to the host computer database and generates test results.

[0010] Preferably, when the key test module performs a performance test, it includes: The material elastic coefficient and stroke threshold of the current test button are retrieved, and the driving air pressure of the needle cylinder is dynamically adjusted to the corresponding initial test pressure through the PID closed-loop control algorithm; The displacement S of the cylinder piston rod, the reaction force F of the key, and the response delay time T of the FTU during key triggering are collected in real time. Multi-dimensional correlation analysis is performed through the data fusion model to generate the key dynamic response curve. Perform nonlinear alignment between the key dynamic response curve and the standard curve, calculate the similarity score between the two, and determine the response performance; Among them, the extreme points of the first-order derivative of the key dynamic response curve are extracted to identify the risk of key sticking.

[0011] Preferably, generating a key dynamic response curve includes: The data fusion model receives normalized time series data of displacement S, reaction force F, and response delay time T through the input layer, extracts local features through a one-dimensional convolution with a convolution kernel size of 3×3 in the first hidden layer, captures long-term dependencies using a bidirectional LSTM in the second hidden layer, and dynamically allocates feature weights for the contact, trigger, and rebound test stages through the attention output layer. The final output is a key dynamic response curve including the trigger threshold point, peak force point, midpoint of the stable response segment, rebound starting point, and reset end point. Among them, during the model training stage, perturbations are added to the S, F, and T data through the fast gradient sign method to generate adversarial samples, and the dual loss function is used to optimize the model.

[0012] Preferably, when the short-circuit pin test module performs a performance test, the following steps are included: The short-circuit pin test module uses a servo motor to drive a precision ball screw to achieve pin feed, collects the contact resistance value between the pin and the FTU interface in real time, and dynamically adjusts the feed speed based on a preset resistance threshold curve; When the contact resistance drops to the set value for the first time, the trigger pin pauses to eliminate contact bounce, and then continues to feed to the preset depth and maintains constant pressure to perform a short circuit test; If abnormal fluctuation occurs during the test, reverse retraction is started and the insertion process is re-executed; After the test is completed, the contact resistance time series data is feature extracted through the impedance spectrum analysis algorithm to generate the contact resistance fluctuation coefficient, which is then synchronously stored in the control system database.

[0013] Preferably, when the network port test module performs a performance test, it includes: A servo motor is used to drive the network port test plug to feed along the Z axis. During the feeding process, the insertion force Fz and displacement Sz are collected in real time through the tension and pressure sensor to establish the Fz-Sz dynamic characteristic curve; When the insertion force Fz reaches the preset contact threshold F0, link negotiation of the first and second rate modes is performed in sequence, and the negotiation success rate and negotiation time are recorded; In the second rate mode, the built-in network analysis module injects a pseudo-random binary sequence to continuously monitor the bit error rate within a set time period and simultaneously collect the eye diagram parameters of the network port differential signal pair; If the bit error rate or eye diagram parameters do not meet the threshold requirements, the piezoelectric ceramic driver applies displacement compensation to the network port test plug, and the link negotiation and signal test are re-executed after each adjustment; After the test is completed, a comprehensive scoring report is generated and stored in the control system database.

[0014] Preferably, the DB9 test module performs the performance test including: The high-precision resistance measurement module uses a matrix scanning method to sequentially detect the contact resistance of each pin of the DB9 interface. Simultaneously, a high-precision differential probe is used to collect the transmission delay and signal rise time of the TX and RX signal pairs. When the contact resistance or signal rise time deviates from the set threshold, the piezoelectric ceramic driver applies displacement compensation to the DB9 interface and re-measures. During the electrical performance test phase, a simulated operating current is injected into the DB9 interface through a programmable load module to monitor the voltage drop and overcurrent protection response time in real time. Perform RS-232 and RS-485 protocol handshake tests in sequence, send data packets containing random check bits, record the protocol negotiation success rate and data transmission error rate, and generate a protocol compatibility matrix report; Among them, the collected multi-dimensional feature data is normalized, the anomaly score of each sample is calculated through the isolation forest model, abnormal samples are identified, and abnormal parameters are extracted.

[0015] Preferably, when the aviation plug test module performs a performance test, it includes: The contact pressure, pin temperature and high-frequency interference current during the plugging process are collected synchronously through the micro pressure sensor, temperature sensor and high-frequency current probe integrated in the aviation plug test module; When the contact pressure fluctuates or the temperature exceeds the limit, the pressure compensation mechanism based on the fuzzy PID algorithm is activated to dynamically adjust the feed rate of the dual-axis servo motor, and the EMI filter is used to suppress high-frequency interference in real time.

[0016] Preferably, dynamically adjusting the feed rate of the dual-axis servo motor includes: The contact pressure fluctuation, pin temperature change rate and high-frequency interference current amplitude are used as input variables of the fuzzy PID controller; A three-dimensional fuzzy rule base that supports online updating is established. The contact pressure fluctuation, pin temperature change rate, and high-frequency interference current amplitude are divided into multiple fuzzy subsets. The fuzzy output of the PID parameter correction is obtained through the centroid method. When the pressure fluctuation still exceeds the threshold after multiple consecutive compensations, the genetic algorithm is automatically triggered to optimize the membership function of the rule antecedent.

[0017] Preferably, when dynamically adjusting the feed rate of the dual-axis servo motor, the method further includes: Dynamically switch the control mode according to the current pressure range: Incremental PID is used in the coarse adjustment stage, and position PID and feedforward compensation are enabled in the fine adjustment stage; When the high-frequency interference current amplitude exceeds the set value, the pressure feedback signal is noise-suppressed by the Kalman filter, and an attenuation coefficient is applied to the PID output; After each feed rate adjustment, the pressure fluctuation standard deviation is continuously monitored. When the pressure fluctuation standard deviation is stable below the set value for consecutive sampling periods, the compensation is considered complete. Otherwise, the parameter self-tuning process is re-executed.

[0018] Compared with the prior art, the present invention has the following beneficial effects: (1) Test accuracy and reliability are significantly improved The coordinated positioning of a two-stage floating locating pin and a dual-V-groove guide block improves interface docking accuracy, reduces contact pressure fluctuations, and resolves interface misalignment issues caused by traditional rigid positioning. The core test unit utilizes PID closed-loop control and a data fusion model to achieve high-precision quantitative analysis of key dynamic response curves, network port eye diagram parameters, and contact resistance, significantly reducing the missed detection rate of hidden faults compared to traditional continuity testing.

[0019] (2) Testing efficiency and automation level have been greatly improved The carrier board can be quickly replaced (with a quick-release mechanism featuring hand-tightened screws) and automatic positioning compensation, shortening changeover time and adapting to the testing needs of multiple FTU models and mixed lines. Fully automated testing (including safety triggering, multi-module collaborative pressing, and data upload) shortens the test cycle and significantly improves efficiency compared to manual intervention. Intelligent compensation mechanisms (such as network port piezoelectric ceramic fine-tuning and aviation plug fuzzy PID pressure compensation) automatically correct abnormal parameters, reducing manual debugging costs.

[0020] (3) Enhanced safety protection and environmental adaptability A three-level safety self-test (optical path calibration, ambient light adaptive threshold, and interlock communication test) combines dual-wavelength collaborative detection (infrared + visible light) with mechanical limit confirmation to ensure the safety light curtain's response time and reduce false triggering. An acrylic protective cover and EMI filter create an electromagnetic shielding environment, improving high-frequency interference suppression and ensuring test data stability in complex industrial scenarios.

[0021] (4) Intelligent data management and fault diagnosis The PLC module synchronizes test data (including dynamic response curves, eye diagram parameters, and protocol negotiation logs) with the host computer database in real time, generating a comprehensive, time-stamped score report. The isolation forest model detects anomalies in multi-dimensional feature data (contact resistance, bit error rate, and protocol success rate), achieving high accuracy in identifying anomalies and automatically correlating production batch information to provide data support for production line process optimization.

[0022] (5) Comprehensive upgrade of working condition simulation and compatibility verification The programmable load module enables 0-500mA dynamic current simulation, accurately replicating actual operating conditions such as motor start-stop and sensor sampling, and minimizing overcurrent protection response time testing errors. It supports RS-232 / RS-485 dual-protocol handshake and 100BASE-TX / 1000BASE-T dual-rate network port testing, covering different FTU interface types and communication scenarios, significantly increasing the compatibility verification dimension compared to traditional solutions.

[0023] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a step diagram of the FTU dynamic function testing method according to an embodiment of the present invention; Figure 2 This is a flow chart of entering a real-time human proximity detection state according to an embodiment of the present invention; Figure 3 This is a flow chart of the performance test performed by the key test module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0026] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0027] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present disclosure, its application, or uses.

[0028] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0029] Example 1, a high-precision multi-interface coordinated FTU dynamic function test system, includes: a protection unit, a test carrying unit, a core test unit, a safety detection unit and a control system.

[0030] The protection unit consists of an acrylic transparent protective cover and a sheet metal box. The acrylic transparent protective cover covers the sheet metal box to form a test space. The sheet metal box is integrated with a switch panel, a pressure gauge, an ammeter and a 24V power module.

[0031] The test carrier unit includes a quickly replaceable carrier plate, an access structure, and a positioning assembly. The carrier plate is fixed to the slide rail via a quick-release structure with hand-tightened screws. The access structure is driven by a rodless cylinder to achieve the in and out movement of the carrier plate. The positioning assembly includes positioning pins and guide blocks for precise positioning of the FTU product to be tested.

[0032] The core test unit includes a pressure plate structure and a test module. The pressure plate structure is driven by a three-axis cylinder to achieve lifting and pressing actions. The test module is located under the pressure plate structure and integrates a button test module, a short-circuit pin test module, a network port test module, a DB9 test module and an aviation plug test module. The button test module is driven by a needle-type cylinder.

[0033] The safety detection unit includes a safety light grid, an anti-pinch device and a pre-feeding sensing mechanism. The safety light grid is set at the entrance of the test space to detect human approach in real time. The pre-feeding sensing mechanism is used to scan and confirm whether the product lid is in the open state.

[0034] The control system is electrically connected to the test carrier unit, core test unit and safety detection unit to control the start-up, execution and result uploading of the test process. The control system is equipped with a PLC module and database storage function.

[0035] Among them, the core test unit also includes an endoscope, which is arranged under the pressure plate structure to collect image data during the test in real time and feed it back to the control system.

[0036] In a possible embodiment, the test module also includes a rotary switch test module and a lever structure. The lever structure controls the switching of the toggle switch through a dual-axis lever cylinder drive, and the rotary switch test module controls the switching of the rotary switch through a joint drive of a rotary cylinder and a blocking cylinder.

[0037] In a possible embodiment, the pre-feeding sensing mechanism includes: a photoelectric sensor and a mechanical limit switch; wherein the photoelectric sensor is used to identify the light-shielding characteristics of the folded edge of the lid for non-contact preliminary judgment; and the mechanical limit switch is used for physical contact with the rear end surface of the folded lid for secondary confirmation.

[0038] In a possible embodiment, a light sensor is integrated on the top of the acrylic transparent protective cover to collect real-time light intensity in the test space.

[0039] In one possible embodiment, the positioning pin adopts a two-section floating structure, consisting of a front guide section and a rear elastic compensation section. The guide section is set to a cone with a taper of 1:12, and the surface is covered with a polytetrafluoroethylene anti-friction coating; the elastic compensation section has a built-in disc spring group, which allows a displacement compensation of ±0.3mm in the axial direction, and is integrated with a micro pressure sensor for real-time monitoring of the positioning contact pressure.

[0040] In a possible embodiment, the guide block is configured as a double V-groove structure, the groove width tolerance is controlled at ±0.02mm, a ceramic wear-resistant strip is embedded in the groove, and the bottom of the guide block is connected to the fine-tuning screw through an eccentric cam to achieve precise adjustment of ±0.1mm in the X and Y axis directions.

[0041] In a possible embodiment, the short-circuit pin test module includes: a servo motor, a precision ball screw, and a pin; wherein the pin is fed by driving the precision ball screw through the servo motor.

[0042] In one possible embodiment, the network port test module includes: a servo motor, a pull-pressure sensor, a built-in network analysis module, a piezoelectric ceramic driver, and a network port test plug; the servo motor is used to drive the network port test plug to feed along the Z axis; the pull-pressure sensor is used to collect the insertion force Fz (sampling frequency ≥ 500 Hz) and displacement Sz in real time; the built-in network analysis module is used to inject a pseudo-random binary sequence (PRBS9) and continuously monitor the bit error rate (BER) within 10 seconds; and the piezoelectric ceramic driver applies displacement compensation of ±0.05 mm in the X and / or Y axis directions to the network port test plug.

[0043] In one possible embodiment, a DB9 test module includes: a DB9 interface, a high-precision resistance measurement module, a high-precision differential probe, and a piezoelectric ceramic driver; wherein the high-precision resistance measurement module is used to sequentially detect the contact resistance of each pin of the DB9 interface using a matrix scanning method; the high-precision differential probe is used to collect the transmission delay and signal rise time of the TX and RX signal pairs; and the piezoelectric ceramic driver is used to apply displacement compensation of ±0.02 mm in the X and / or Y axis directions to the DB9 interface.

[0044] In a possible embodiment, the DB9 test module further includes: a programmable load module, which is used to inject simulated working current into the DB9 interface and monitor the voltage drop and overcurrent protection response time in real time.

[0045] In one possible embodiment, the aviation plug test module includes: a micro pressure sensor, a temperature sensor, a high-frequency current probe, an EMI filter, a fuzzy PID controller, and a dual-axis servo motor; wherein the micro pressure sensor, temperature sensor, and high-frequency current probe are respectively used to synchronously collect contact pressure, pin temperature, and high-frequency interference current during the plugging process; the EMI filter is used to suppress high-frequency interference in real time; the fuzzy PID controller uses the contact pressure fluctuation, the pin temperature change rate, and the high-frequency interference current amplitude as input variables to dynamically adjust the feed amount of the dual-axis servo motor.

[0046] Example 2, see Figure 1 The present invention provides a step diagram of the FTU dynamic function test method. Figure 1 A high-precision multi-interface collaborative FTU dynamic function test method is shown, comprising the following steps: Step S1, device startup and status confirmation: The protective unit consisting of the acrylic transparent protective cover and the sheet metal box is energized and ventilated, and the pressure gauge, ammeter and 24V power module integrated in the sheet metal box are used to confirm that the initial status of the device is normal; Step S2, product loading and positioning preparation: adjust and fix the carrier plate by using the quick-release mechanism of the carrier plate's hand screws, place the FTU product to be tested on the carrier plate, manually open the product cover, and trigger the pre-feed sensor mechanism to scan and confirm that the cover is in the open state; Step S3, dual start and safety trigger: Press the left and right start buttons on the sheet metal box at the same time, and the control system starts the safety detection unit, so that the safety grating set at the entrance of the test space is powered on and enters the real-time human proximity detection state; Step S4: Carrier positioning and test position entry: The carrier is driven by the rodless cylinder of the test carrier unit to move along the slide rail, and the positioning pins and guide blocks of the positioning assembly cooperate to accurately position the product, and the carrier enters the test position of the core test unit; Step S5, multi-module collaborative press fit test: The control system drives the three-axis cylinder of the core test unit to drive the pressure plate structure downward, so that the button test module, short-circuit pin test module, network port test module, DB9 test module, and aviation plug test module integrated in the pressure plate structure sequentially contact the interface of the FTU product to be tested to perform performance testing; Step S6, data acquisition and result upload: During the test process, the endoscope collects test image data in real time, and the control system synchronously uploads the test data and image data to the host computer database and generates the test results; Step S7, safe unloading and cycle preparation: After the test is completed, the three-axis cylinder drives the pressure plate structure to rise, the rodless cylinder pushes the carrier plate to the initial position, the anti-pinch device is activated and a material removal prompt is issued, completing a single test cycle; In a possible embodiment, when performing the performance test, the following steps are further included: The dual-axis lever cylinder drives the lever structure to control the switching of the toggle switch, and the rotary cylinder and the blocking cylinder jointly drive the rotary switch test module to control the switching of the rotary switch.

[0047] Background description: In the automated testing process, confirming the initial state of the product to be tested is a key prerequisite to ensure the safety and accuracy of the test. For the FTU functional test system, if the product cover is not in an effectively open state, it may cause abnormal docking between the test module and the product interface, causing the risk of mechanical collision or electrical signal misjudgment. Traditional detection methods often rely on a single sensor (such as a photoelectric sensor or a mechanical limit switch), which is easily affected by factors such as ambient light interference, differences in the reflective properties of the product surface, or wear of mechanical contacts, and there is a risk of false detection or missed detection. Based on this: In the above step S1, when scanning to confirm that the cover is in the open state, the following steps are performed: A dual detection mechanism employs a photoelectric sensor and a mechanical limit switch. The photoelectric sensor identifies the light-shielding characteristics of the lid's folded edge for a non-contact initial determination, while the mechanical limit switch provides secondary confirmation through physical contact with the lid's rear end. The system determines the lid is effectively opened only when, and only when, the photoelectric sensor detects the disappearance of the light-shielding signal for three consecutive sampling times (with a 0.5-second sampling interval) and the mechanical limit switch triggers a closing signal. This generates a timestamped status confirmation log and stores it in the control system database. Among them, the photoelectric sensor adopts a dual-wavelength collaborative detection mechanism: it synchronously emits infrared light (wavelength 850nm) and visible light (wavelength 550nm), and collects the reflected light intensity of the two wavelengths through the receiver respectively; sets the dynamic ratio coefficient of the infrared light shading threshold and the visible light shading threshold (range 1.2-1.5). When the reflected light intensity of the two wavelengths is lower than their respective thresholds in three consecutive samplings (sampling interval 0.5 seconds) and the dynamic ratio coefficient is within the set range, it is determined that the shading signal has effectively disappeared.

[0048] What needs further explanation in the embodiments of the present invention is that this embodiment designs a dual detection mechanism and dual-wavelength collaborative verification logic. By combining the use of photoelectric sensors and mechanical limit switches, the dual protection of non-contact preliminary judgment and physical contact secondary confirmation is achieved; at the same time, the photoelectric sensor uses dual-wavelength emission and reception of infrared light (850nm) and visible light (550nm), combined with a dynamic ratio coefficient algorithm, to effectively suppress false triggering that may occur under a single wavelength (such as dust obstruction, changes in material transmittance, etc.). This design not only meets the needs of automated testing for rapid response, but also improves the robustness of state judgment through multi-dimensional verification.

[0049] Background description: During the operation of the automated test system, the core test area involves mechanical movements such as the pressure plate structure driven by a three-axis cylinder and the carrier movement driven by a rodless cylinder. If a person's limbs accidentally enter during the test execution phase, it may cause mechanical collision or extrusion risks. Traditional safety protection mechanisms often have problems such as response delays, insufficient environmental adaptability, or imperfect self-test mechanisms. For example, optical sensors are easily disturbed by changes in light and may be triggered incorrectly, or communication failures in safety interlock devices may not be discovered in time, thus posing safety hazards. Based on this: See Figure 2 The flow chart of entering the real-time human proximity detection state, in the above step S3, when entering the real-time human proximity detection state, includes: When the safety light curtain is started, a three-level self-test process is executed: optical path calibration, signal threshold configuration and linkage equipment communication test are completed in sequence; Among them, optical path calibration: the transmitter lights up each light beam in sequence at a preset frequency, and the receiver records the light intensity attenuation value of each light beam. When the attenuation value exceeds the threshold (5%), the driving current is automatically compensated; Signal threshold configuration: Adopting an ambient light adaptive algorithm, the light shading detection threshold is dynamically adjusted based on the real-time light intensity in the test space (collected by the light sensor integrated on the top of the acrylic transparent protective cover). For every 1000 lux increase in light intensity, the threshold increases by 8%. Communication test of linked equipment: Send a simulated trigger signal to the anti-pinch device through the PLC module to verify the safety interlock response time (required to be ≤20ms). If the self-test is abnormal, a fault code will be generated and the test process will be locked until it is manually reset.

[0050] Further explanation is required regarding the embodiment of the present invention, which incorporates a three-level self-check process for real-time human proximity detection. Optical path calibration ensures the stability of the grating's transmitted and received light beams, avoiding detection blind spots due to light intensity attenuation. An ambient light adaptive algorithm dynamically adjusts the threshold to address the impact of lighting changes within the test space (such as fluctuations in external light passing through the acrylic protective cover) on detection accuracy. Linked device communication testing verifies the response timeliness of interlocking components such as the safety grating and anti-pinch device, ensuring that the device can trigger shutdown protection within 20ms or less in an emergency. This design, through a multi-layered protection logic of "hardware self-check + environmental adaptation + interlock verification," improves the reliability of human proximity detection while reducing production efficiency losses due to false detections.

[0051] Background description: In the automated testing process, precise docking of the FTU product to be tested and the test module is the core prerequisite for ensuring the accuracy of interface test data. Traditional positioning mechanisms often use fixed pin-hole matching or single-dimensional adjustment structures, which are easily affected by factors such as product processing tolerances (such as housing positioning hole deviation), cumulative errors in carrier board installation, and wear of mechanical components. When the positioning deviation exceeds 0.1mm, problems such as misalignment between the test plug and the product interface, poor contact, and even physical damage may occur. In addition, the process of manually adjusting positioning parameters is cumbersome and requires repeated calibration, which seriously restricts test efficiency. Based on this: In the above step S4, when accurately positioning the product, it includes: The locating pin adopts a two-stage floating structure, consisting of a front guide section and a rear elastic compensation section. The guide section is a cone with a taper of 1:12 and is covered with a polytetrafluoroethylene anti-friction coating. The elastic compensation section has a built-in disc spring group, allowing for an axial displacement compensation of ±0.3mm. It also integrates a micro pressure sensor for real-time monitoring of the positioning contact pressure (threshold range 5-8N). The guide block is set to a double V-groove structure with a groove width tolerance controlled at ±0.02mm. Ceramic wear-resistant strips are embedded in the groove, and the bottom of the guide block is connected to the fine-tuning screw through an eccentric cam to achieve precise adjustment of ±0.1mm in the X and Y axis directions; During the positioning process, the control system drives the fine-tuning screw to automatically compensate for positioning deviation by comparing the feedback value of the micro pressure sensor with the preset threshold. When the pressure fluctuation value of three consecutive samples (with an interval of 0.2 seconds) is ≤0.5N, the positioning is determined to be completed and a positioning log with coordinate parameters is generated.

[0052] Further explanation is required regarding this embodiment of the present invention. To meet the high-precision requirements for simultaneous testing of multiple interfaces (such as network ports, DB9 connectors, and aviation plugs) on FTU products, this embodiment utilizes a dual-stage floating locating pin and a dual-V-groove guide block for coordinated positioning. The 1:12 tapered guide section at the front of the locating pin, coated with polytetrafluoroethylene, enables adaptive guidance during insertion. A disc spring assembly at the rear provides ±0.3mm axial compensation, preventing stress concentration caused by rigid contact. The dual-V-groove guide block utilizes ceramic wear strips and an eccentric cam for fine-tuning, enabling precise adjustment of the X and Y axes to ±0.1mm, accommodating housing tolerances across product batches. Furthermore, a micro-pressure sensor monitors contact pressure in real time (with a threshold range of 5-8N). Combined with a dynamic compensation algorithm, this ensures pressure fluctuations of ≤0.5N during positioning, meeting the mechanical precision requirements (±0.02mm) for interface docking while preventing product deformation caused by over-positioning. Compared to traditional positioning methods, this design reduces commissioning time and improves positioning repeatability, providing a foundation for consistent electrical performance during subsequent multi-module collaborative testing.

[0053] Background description: In the functional testing of FTU products, buttons are key components of human-computer interaction. Their mechanical response characteristics (such as trigger force, travel, and rebound speed) and electrical response delay directly affect the product experience and reliability. Traditional button testing mostly uses fixed air pressure drive or mechanical limit methods, which have the following limitations: First, it is impossible to dynamically adjust the trigger parameters according to the differences in the elastic coefficients of the button materials (such as silicone and metal domes), which can easily lead to overvoltage damage or poor contact; second, the qualification is judged only by conductivity testing, ignoring hidden defects such as jamming and slow response during the dynamic process of the button; third, the manual recording of response time has large errors, and it is difficult to establish a correlation analysis model for force-displacement-time multidimensional data. Based on this: See Figure 3 In the flowchart of the key test module performing the performance test, in the above step S5, when the key test module performs the performance test, it includes: Before triggering a key with a needle-type cylinder, the key test module retrieves the material elastic coefficient and stroke threshold of the current test key from the preset key characteristic database. Based on the elastic coefficient and stroke threshold, the driving air pressure of the needle-type cylinder is dynamically adjusted to the corresponding initial test pressure P0 through a PID closed-loop control algorithm. During the key triggering process, the displacement S of the cylinder piston rod, the key reaction force F, and the response delay time T of the FTU are collected in real time. When it is detected that the key reaction force F exceeds the preset safety threshold Fmax, the cylinder pressure relief protection mechanism is immediately triggered and the fault data is recorded; After the test is completed, a multi-dimensional correlation analysis of S, F, and T is performed through the data fusion model to generate a key dynamic response curve; The dynamic time warping (DTW) algorithm is used to nonlinearly align the key dynamic response curve with the standard curve, and the similarity score between the two is calculated to determine the response performance (if the similarity score is greater than the threshold, it is preliminarily determined to be qualified); Among them, the extreme points of the first-order derivative (dF / dS) of the key dynamic response curve are extracted. If ≥2 local maximum values ​​appear in the trigger stage, it is determined that there is a risk of key sticking.

[0054] What needs to be further explained in the embodiments of the present invention is that this embodiment matches the elastic coefficient and stroke threshold of different product models through a preset key characteristic database, and combines the PID closed-loop control algorithm to adjust the needle cylinder driving air pressure in real time to achieve "soft contact" triggering; synchronously collects the piston rod displacement (S), key reaction force (F) and FTU response delay time (T), adopts an improved LSTM deep learning model to fuse multi-dimensional time series data, constructs a dynamic response curve, and compares it with the standard curve through the DTW algorithm to achieve multi-dimensional evaluation of key response performance, and combines the first-order derivative extreme point analysis to accurately identify the risk of sticking (such as the appearance of ≥2 local maximum values ​​in the triggering stage).

[0055] Background description: There are three core limitations in the traditional method of generating dynamic response curves of buttons: First, the ability to fuse multi-source time series data (displacement S, reaction force F, response delay time T) is insufficient, and simple weighting or feature splicing methods are often used, resulting in the smoothing of key dynamic features (such as sudden force changes in the trigger phase); second, the model has weak generalization ability and is sensitive to interference factors such as sensor noise and oil stains on the button surface, which can easily cause curve distortion; third, the test phase is vaguely divided, and the feature weights of different phases such as contact, triggering, and rebound are not differentiated, which affects the extraction accuracy of key feature points of the curve (such as peak force point and rebound starting point). Based on this: In a possible embodiment, generating a key dynamic response curve includes: The data fusion model uses a deep learning architecture based on an improved long short-term memory (LSTM) network. The deep learning architecture consists of an input layer, two hidden layers, and an attention output layer. The input layer receives normalized time series data of displacement S, reaction force F, and response delay time T. The first hidden layer extracts local features through one-dimensional convolution with a convolution kernel size of 3×3. The second hidden layer uses a bidirectional LSTM to capture long-term dependencies. The attention output layer dynamically assigns feature weights for different test phases (contact / trigger / rebound), ultimately outputting a dynamic response curve for the keystroke containing five feature points: trigger threshold point, peak force point, midpoint of the stable response segment, rebound starting point, and reset end point. During the model training phase, FGSM (Fast Gradient Signed Method) is used to add perturbations (perturbation amplitude ≤ 3%) to the S, F, and T data to generate adversarial samples. The model is optimized using a dual loss function: the main loss function is the mean square error (MSE), and the auxiliary loss function is the curve deviation rate between the adversarial sample and the original sample; After training, the model's tolerance for abnormal data (such as sudden noise from sensors and F-value jumps caused by oil on the button surface) is improved, and the misjudgment rate is significantly reduced.

[0056] The present invention further explains its proposed deep learning architecture based on an improved LSTM. A dual-hidden-layer design (one-dimensional convolution + bidirectional LSTM) achieves the coordinated capture of local dynamic features and long-term dependencies. An attention mechanism dynamically weights the contribution of features at different test stages, ensuring the accuracy of key nodes such as the trigger threshold. The innovative FGSM adversarial sample training injects perturbations of up to 3% into the original data, improving the model's tolerance to abnormal data. A dual loss function (mean squared error (MSE) + adversarial sample bias rate) is used to optimize network parameters, significantly reducing curve fitting error. The resulting dynamic response curve contains five key feature points, providing a quantitative basis for key sticking risk (based on first-order derivative extreme point analysis) and performance consistency assessment.

[0057] Background description: In traditional short-circuit pin tests, the stability of contact resistance and feed control accuracy are the core bottlenecks that restrict test reliability. Traditional drive methods mostly use cylinders or electromagnetic push rods, which have problems such as unadjustable feed speed and large fluctuations in contact pressure between the pin and the interface (more than ±2N). This causes the contact resistance to easily jump by >50mΩ during the conduction stage. At the same time, there is a lack of a real-time resistance monitoring mechanism, and it is impossible to identify poor contact caused by contact oxidation or foreign matter intervention. The result can only be judged through post-conductivity testing, with a hidden fault missed detection rate of about 3%. In addition, the mechanical bounce during the pin insertion process (duration 0.1-0.5 seconds) will cause current waveform distortion, affecting the verification accuracy of the short-circuit protection algorithm. Based on this: In the above step S5, when the short-circuit pin test module performs the performance test, it includes: A dual-mode latch drive and contact resistance dynamic monitoring mechanism is used: the short-circuit latch test module uses a servo motor to drive a precision ball screw to achieve latch feed. During the feed process, the contact resistance value between the latch and the FTU interface is collected in real time (sampling frequency ≥ 1kHz). The feed speed is dynamically adjusted (range 0.5-2mm / s) based on a preset resistance threshold curve (initial contact stage threshold ≤ 50mΩ, stable conduction stage threshold ≤ 10mΩ); When the contact resistance drops to 50mΩ for the first time, the trigger pin pauses for 0.3 seconds to eliminate contact bounce, then continues to feed to the preset depth (accuracy ±0.01mm) and maintains a constant pressure (3-5N) to perform a short circuit test; During the test, if the contact resistance fluctuates by ≥5mΩ during the stable conduction phase, or if the number of fluctuations exceeds 3 times within a single test cycle, the system automatically initiates reverse retraction (retraction distance 0.2mm) and re-executes the insertion process. The cumulative number of retries shall not exceed 2 times. After the test is completed, the contact resistance time series data is feature extracted through the impedance spectrum analysis algorithm to generate the contact resistance fluctuation coefficient, which is then synchronously stored in the control system database.

[0058] Further explanation is required regarding the embodiment of the present invention. This embodiment utilizes a servo motor and ball screw combination to achieve stepless adjustment of the feed speed (0.5-2 mm / s), coupled with 1kHz high-frequency resistance sampling to create a closed-loop feedback system. An innovative three-stage insertion process, consisting of "threshold triggering - pause debounce - constant pressure hold," is employed. When the contact resistance first drops to 50mΩ, a 0.3-second pause is triggered, mechanical damping eliminates bouncing interference, and contact stability is subsequently maintained with a constant pressure of 3-5N. An adaptive retry mechanism (maximum two retraction and re-insertion cycles) is developed to address resistance fluctuations during the conduction phase, ensuring a high repair rate for contact failures in a single test. Finally, impedance spectrum analysis is used to extract the time-domain characteristics of the contact resistance (such as the fluctuation coefficient), providing data support for pin wear life prediction. This mechanism improves contact resistance stability and reduces short-circuit current waveform distortion, significantly enhancing the reliability of FTU overcurrent protection testing.

[0059] Background description: In the traditional network port testing process, there are obvious technical bottlenecks in the coordinated verification of mechanical connection reliability and signal transmission quality. Traditional tests mostly use cylinders or electromagnetic drives to plug and unplug plugs, which have three core problems: First, the plugging and unplugging force is uncontrollable, which can easily cause deformation of the network port structure due to overpressure (such as bending of the crystal head pins) or insufficient contact pressure to cause signal attenuation, and it is impossible to record the force-displacement dynamic characteristics during the insertion process; second, the rate compatibility verification is not comprehensive, usually only testing a single rate mode (such as 100BASE-TX), ignoring the link negotiation efficiency and stability in high-speed modes such as 1000BASE-T; third, the signal integrity assessment dimension is limited, and only basic conduction tests are used to judge eligibility. There is a lack of quantitative analysis of key indicators such as bit error rate (BER) and eye diagram parameters, making it difficult to detect hidden defects such as high-frequency interference and impedance mismatch. In addition, when parameter abnormalities occur during the test process, the traditional system requires manual adjustment of the plug position, which is inefficient and has poor repeatability, and cannot meet the high-precision testing requirements of automated production lines. Based on this: In the above step S5, when the network port test module performs the performance test, it includes: A servo motor is used to drive the network port test plug to feed along the Z axis. During the feeding process, the insertion force Fz (sampling frequency ≥ 500Hz) and displacement Sz are collected in real time through the tension and pressure sensor to establish the Fz-Sz dynamic characteristic curve; When the insertion force Fz reaches the preset contact threshold F0 (3-5N), the network port communication initialization process is triggered: link negotiation of the two speed modes of 100BASE-TX and 1000BASE-T is performed in sequence, and the negotiation success rate and negotiation time (threshold ≤ 100ms) are recorded; In 1000BASE-T mode, the built-in network analysis module injects a pseudo-random binary sequence (PRBS9) and continuously monitors the bit error rate (BER) within 10 seconds, requiring BER ≤ 1e-12. Synchronously collect eye diagram parameters of the network port differential signal pair (TX+ / TX-, RX+ / RX-), including eye height (≥800mV), eye width (≥0.8UI), and crossing point jitter (≤100ps); If the bit error rate (BER) or eye diagram parameters do not meet the threshold requirements, the system automatically activates the fine-tuning mechanism: the piezoelectric ceramic driver applies a displacement compensation of ±0.05mm in the X and / or Y axis direction to the network port test plug. After each adjustment, the link negotiation and signal test are re-executed. The cumulative number of adjustments shall not exceed three times. After the test is completed, a comprehensive scoring report is generated that includes mechanical connection characteristics (Fz-Sz curve similarity ≥ 95%), electrical performance (BER, eye diagram parameters), negotiation success rate and negotiation time, and stored in the control system database.

[0060] What needs to be further explained in the embodiment of the present invention is that the network port test module achieves high-precision testing through a three-level collaborative mechanism of "mechanical characteristics-electrical performance-adaptive compensation". Mechanically, a servo motor drives the ball screw, coupled with a pull-and-load sensor (sampling frequency ≥ 500Hz) to collect insertion force (Fz) and displacement (Sz) in real time. This generates an Fz-Sz dynamic characteristic curve to ensure the stability of physical contact between the plug and the network port (contact threshold 3-5N). Electrical performance testing is conducted in two phases: First, basic communication capabilities are verified through 100BASE-TX / 1000BASE-T dual-rate link negotiation (negotiation time ≤ 100ms). A PRBS9 pseudo-random sequence is then injected to monitor the bit error rate (BER ≤ 1e-12) for 10 seconds. Eye diagram parameters of the differential signal pair are simultaneously collected (eye height ≥ 800mV, eye width ≥ 0.8UI, jitter ≤ 100ps). To address performance degradation caused by mechanical deviations, the system features a built-in piezoelectric ceramic fine-tuning mechanism that allows for ±0.05mm displacement compensation in the X and Y axes (up to three iterations). Finally, a comprehensive report is generated that includes mechanical connection similarity (≥95%), electrical parameters, and negotiation time, enabling coordinated verification of physical contact and signal integrity.

[0061] Background description: There are two core technical limitations in the traditional DB9 interface test process: First, the conductivity and signal integrity tests are separated from each other. Independent processes are often used to detect pin contact resistance and signal transmission parameters respectively, ignoring the implicit impact of poor contact on high-frequency signals (such as TX and RX differential pairs) (such as signal attenuation caused by increased contact resistance); second, there is a lack of adaptive compensation mechanism. When excessive contact resistance or signal distortion is detected, the interface position needs to be manually adjusted, resulting in long debugging cycles and low positioning accuracy, making it difficult to meet the test efficiency and consistency requirements of automated production lines. Based on this: In the above step S5, the DB9 test module performs the performance test, including: Collaborative testing of pin-independent continuity and differential signal integrity: A high-precision resistance measurement module uses a matrix scanning method to sequentially test the contact resistance of each pin of the DB9 interface (threshold ≤ 15mΩ). Simultaneously, a high-precision differential probe is used to collect the transmission delay (≤ 30ns) and signal rise time (10%-90% amplitude range threshold ≤ 20ns) of the TX and RX signal pairs. When the contact resistance of any pin exceeds the threshold or the signal rise time deviates from the standard value by ±20%, the pin position fine-tuning mechanism is triggered: the piezoelectric ceramic driver is used to apply a displacement compensation of ±0.02mm in the X and / or Y axis direction to the DB9 interface, and the parameters are re-tested until they meet the standards (cumulative adjustment times ≤ 3 times).

[0062] It is worth further explaining that in the embodiments of the present invention, the matrix scanning method can quickly locate pins with poor contact (such as abnormal resistance caused by oxidation or misalignment), avoiding the low efficiency of traditional single-channel detection. While conducting the continuity test, a high-precision differential probe is used to access the TX and RX signal pairs, collecting two sets of key parameters in real time: Transmission delay: measures the time difference between the signal sending end and the receiving end, which must be ≤30ns to ensure the real-time data transmission; Signal rise time: Based on the 10%-90% amplitude range, the threshold is ≤20ns. It reflects the steepness of the signal edge and avoids signal distortion caused by impedance mismatch.

[0063] A piezoelectric ceramic driver applies displacement compensation to the DB9 interface in the X and / or Y axis directions. Continuity and signal tests are re-performed after each adjustment, with a cumulative maximum of three adjustments to ensure the optimal connection within mechanical tolerances.

[0064] Conductivity and signal integrity testing are performed simultaneously, avoiding the hidden risk of "continuity passed but signal distortion" in traditional step-by-step testing. For example, excessive contact resistance can directly lead to increased signal transmission delay, while abnormal rise time can infer insufficient pin contact pressure. Cross-validation of these two data points can improve fault location accuracy.

[0065] This principle deeply couples mechanical connection quality with electrical performance indicators through the process of "real-time monitoring-dynamic compensation-closed-loop verification", ensuring the communication reliability of the DB9 interface under complex working conditions.

[0066] Background description: In traditional DB9 interface testing, there are significant technical bottlenecks in the verification of dynamic working condition adaptability and protocol robustness: First, the load simulation is static, and fixed current values ​​(such as 100mA or 200mA) are mostly used for testing. It is impossible to reproduce the dynamic current fluctuation scenarios of 0-500mA in actual applications (such as motor start and stop, sensor burst sampling), resulting in distortion of key parameter tests such as overcurrent protection response time (required ≤100ms); second, the protocol test is single, usually only verifying the basic RS-232 communication function, ignoring the protocol negotiation efficiency in the RS-485 multi-machine networking scenario, and the data packet has no random check bits, making it difficult to expose compatibility defects under complex working conditions (such as communication interruption caused by check errors); third, abnormal diagnosis is fragmented, and multi-dimensional data such as contact resistance, signal delay, protocol negotiation success rate and data transmission bit error rate are stored independently, lacking normalized fusion analysis. Based on this: In the above step S5, when the DB9 test module performs the performance test, it also includes: Dynamic load simulation and protocol compatibility verification steps: During the electrical performance test phase, a simulated operating current (0-500mA adjustable, 10mA step) is injected into the DB9 interface through a programmable load module to monitor the voltage drop (threshold ≤ 0.3V) and overcurrent protection response time (≤ 100ms) in real time. Perform RS-232 and RS-485 protocol handshake tests in sequence, sending data packets (1024 bytes in length) containing random check bits, recording the protocol negotiation success rate (required to be ≥99.5%) and data transmission bit error rate (BER ≤ 1e-11), and generating a protocol compatibility matrix report; The collected multi-dimensional feature data, such as contact resistance, signal delay, protocol negotiation success rate, and data transmission error rate, are normalized, and the anomaly score of each sample is calculated using the isolation forest model (a threshold of ≥0.7 is considered abnormal). Abnormal samples are ranked by feature importance, and the top three contributing abnormal parameters (such as "contact resistance jump" and "protocol timeout") are extracted. A fault diagnosis report with root cause analysis is generated and automatically associated with the production batch information of the FTU product.

[0067] It is worth further explaining that in the embodiments of the present invention, the dynamic load simulation and protocol compatibility verification steps of the DB9 test module achieve a comprehensive assessment of the actual application capabilities of the interface by combining dynamic working condition reproduction with multi-dimensional protocol robustness verification: A programmable load module injects a continuously adjustable current of 0-500mA (with a step accuracy of 10mA) into the DB9 interface to simulate dynamic current fluctuations in real-world scenarios such as motor start-stop and sensor sudden sampling. Two key parameters are monitored in real time: Voltage drop: Ensures interface power supply stability when the load changes, with a threshold of ≤0.3V, to avoid insufficient power supply to the device due to excessive contact resistance; Overcurrent protection response time: When the current exceeds the safety threshold, the triggering speed of the FTU overcurrent protection mechanism is tested, which is required to be ≤ 100ms to verify its ability to quickly protect against abnormal current.

[0068] Dual-protocol handshake test: The handshake process of RS-232 (point-to-point communication) and RS-485 (multi-machine networking) protocols is executed in sequence, and a 1024-byte data packet containing random check bits is sent. The random check bits are used to simulate data interference in actual communication to verify the fault tolerance of the protocol in complex data scenarios. Key indicators include: Protocol negotiation success rate: ≥99.5% to ensure communication stability when multiple devices are networked; Data transmission bit error rate (BER): The requirement is ≤1e-11, that is, no more than 1 bit error is allowed for every 10^11 bits of data transmitted, to ensure the accuracy of data transmission.

[0069] Multidimensional data anomaly diagnosis: Normalize multi-dimensional feature data such as contact resistance, signal delay, protocol negotiation success rate, and data transmission bit error rate to eliminate dimensional differences; The isolation forest model (an unsupervised anomaly detection algorithm) is used to calculate the sample anomaly score. A threshold of ≥0.7 is considered an anomaly, enabling intelligent identification of hidden faults. Abnormal samples are ranked by feature importance, and the three most contributing abnormal parameters (such as "contact resistance jump" and "protocol timeout") are extracted. Combined with production batch information, a root cause analysis report is generated to provide data support for production line process optimization.

[0070] Background description: In the field of aviation plug testing, traditional testing methods face three core technical challenges: First, the contact pressure control accuracy is insufficient. The use of fixed air pressure or mechanical limit drive can easily cause pressure fluctuations exceeding ±2N due to mechanical tolerances during the plugging process (such as plug shell deformation and pin position deviation), resulting in unstable contact resistance or the risk of pin bending; second, there is a lack of a multi-parameter collaborative monitoring mechanism. The qualification is judged only by conductivity testing, ignoring the implicit impact of temperature rise during the plugging process (such as Joule heat caused by poor contact) on material properties, and the interference of high-frequency electromagnetic interference (1MHz-1GHz) on signal transmission in industrial environments; third, there is a lack of dynamic compensation capability. When the pressure fluctuates or the temperature exceeds the standard, manual intervention and adjustment are required, which cannot meet the requirements of automated production lines for test efficiency and consistency. Based on this: In the above step S5, when the aviation plug test module performs the performance test, it includes: The miniature pressure sensor, temperature sensor, and high-frequency current probe integrated into the aviation plug test module synchronously collect contact pressure (threshold range 8-12N), pin temperature (≤60°C), and high-frequency interference current (frequency 1MHz-1GHz, amplitude ≤50mV) during the plugging process. When the contact pressure fluctuation exceeds ±1N or the temperature exceeds the limit, a pressure compensation mechanism based on the fuzzy PID algorithm is activated to dynamically adjust the feed rate of the dual-axis servo motor (compensation accuracy ±0.005mm), and an EMI filter is used to suppress high-frequency interference in real time to ensure the accuracy of test data in complex electromagnetic environments.

[0071] What needs to be further explained in the embodiment of the present invention is that this embodiment breaks through the limitations of traditional static testing through the closed-loop control of "real-time monitoring-intelligent compensation-interference suppression". It not only verifies the performance of the aviation plug under ideal conditions, but also simulates the dynamic load, temperature changes and electromagnetic interference in actual applications, ensuring the long-term reliability of the interface and signal transmission stability in complex industrial environments.

[0072] Background description: In the traditional dynamic adjustment mechanism of the feed rate of dual-axis servo motors, only contact pressure is used as the feedback signal, ignoring the coupling effect of temperature change rate (dT / dt) and high-frequency electromagnetic interference (I_emi) on system stability, resulting in compensation lag or over-adjustment under complex working conditions (such as sudden temperature rise or strong electromagnetic interference); the rule base relies on offline debugging presets and cannot be dynamically optimized. When the system characteristics drift, the compensation accuracy drops significantly. It is difficult to take into account both fast convergence (coarse adjustment requirements) and steady-state accuracy (fine adjustment requirements) using a single control mode. The traditional adjustment mechanism only judges the compensation effect through a single adjustment result, and does not introduce statistical indicators such as the standard deviation of fluctuation (σ_F), resulting in a certain error compensation rate. Based on this: In a possible embodiment, dynamically adjusting the feed rate of the dual-axis servo motor includes: Real-time input variable construction: The contact pressure fluctuation (ΔFs), pin temperature change rate (dT / dt), and high-frequency interference current amplitude (I_emi) are used as input variables of the fuzzy PID controller. ΔFs is calculated using the differential signal of the micro pressure sensor (sampling frequency ≥ 2kHz), dT / dt is solved in real time using the first-order difference algorithm on the temperature sensor data, and I_emi is obtained by bandpass filtering (1MHz-1GHz) to obtain the effective value. Fuzzy rule dynamic optimization: A three-dimensional fuzzy rule base is established, dividing ΔFs (quantization interval [-2N, +2N]), dT / dt (quantization interval [-5°C / s, +5°C / s]), and I_emi (quantization interval [0mV, 50mV]) into five fuzzy subsets (negative large / negative small / zero / positive small / positive large). The center of gravity method is used to defuzzify and output PID parameter corrections (ΔKp, ΔKi, ΔKd). The rule base supports online updates. If the pressure fluctuation still exceeds the threshold after three consecutive compensations, a genetic algorithm is automatically triggered to optimize the membership function of the rule antecedent (crossover probability 0.6, mutation probability 0.1). Segmented PID control strategy: Dynamically switch the control mode according to the range of the current pressure value (Fs): Coarse adjustment stage (Fs<8N or Fs>12N): adopt incremental PID, aim at fast convergence, Kp takes 0.8-1.2 times the reference value, and Ki is set to zero; Fine-tuning stage (8N≤F≤12N): Enable position PID and feedforward compensation, introduce pressure change trend prediction (based on ARIMA model to model the previous 100ms data), and output 30% of the estimated compensation amount in advance; Anti-interference compensation correction: When I_emi exceeds 30mV, the pressure feedback signal is noise-reduced through the Kalman filter, and an attenuation coefficient (0.7-0.9, positively correlated with I_emi) is applied to the PID output to avoid overcompensation caused by electromagnetic interference; Closed-loop verification of compensation effect: After each feed adjustment (adjustment step ≤ 0.002mm), continuously monitor the standard deviation of pressure fluctuation (σ_F) within 50ms. When σ_F ≤ 0.3N and is stable for three consecutive sampling periods, the compensation is considered complete. Otherwise, the parameter self-tuning process is re-executed (the cumulative number of iterations ≤ 5 times).

[0073] It is necessary to further explain in the embodiment of the present invention that the mechanism of dynamically adjusting the feed rate of the dual-axis servo motor in this embodiment realizes high-precision dynamic compensation during the insertion and insertion of the aviation plug through the collaborative control logic of multi-dimensional perception-intelligent decision-making-closed-loop optimization. The specific principles are as follows: A three-dimensional dynamic sensing system is constructed using the contact pressure fluctuation (ΔFs), the pin temperature change rate (dT / dt), and the high-frequency interference current amplitude (I_emi) as inputs to the fuzzy PID controller: Contact pressure fluctuation (ΔFs): Calculates pressure deviation in real time through the differential signal of the micro pressure sensor (sampling frequency ≥ 2kHz), accurately capturing small fluctuations at the level of ±0.1N, reflecting the contact stability between the plug and the socket; Pin temperature change rate (dT / dt): This uses a first-order difference algorithm to solve temperature sensor data and monitor the pin temperature rise rate (unit: °C / s) in real time, providing early warning of local overheating risks caused by poor contact. High-frequency interference current (I_emi): The effective value is taken after 1MHz-1GHz bandpass filtering to quantify the electromagnetic interference intensity in the industrial environment and avoid false triggering of pressure feedback by interference signals.

[0074] Establish a three-dimensional fuzzy rule base and realize self-iteration of control strategy through dynamic optimization algorithm: Fuzzy subset partitioning: ΔFs ([-2N, +2N]), dT / dt ([-5℃ / s, +5℃ / s]), and I_emi ([0mV, 50mV]) are divided into five fuzzy subsets: negative large / negative small / zero / positive small / positive large, covering the entire range of parameter fluctuations. Defuzzification and PID correction: The center of gravity method is used to convert the fuzzy reasoning results into PID parameter correction values ​​(ΔKp, ΔKi, ΔKd) to achieve dynamic adjustment of proportional, integral and differential coefficients; Online evolution of the rule base: When pressure fluctuations still exceed the threshold after three consecutive compensations, a genetic algorithm is automatically triggered to optimize the membership function of the rule antecedent (crossover probability 0.6, mutation probability 0.1). This simulates the natural selection process to improve the rule base's adaptability to equipment aging and environmental changes.

[0075] The control mode is dynamically switched according to the current contact pressure (Fs) range, taking into account both fast convergence and steady-state accuracy: Coarse adjustment stage (Fs < 8N or Fs > 12N): Incremental PID is used to quickly eliminate large deviations. Kp is set to 0.8-1.2 times the reference value (to enhance proportional action), and Ki is set to zero (to avoid integral saturation) to ensure that the pressure quickly enters the target range. Fine-tuning stage (8N≤Fs≤12N): Enable position PID + feedforward compensation, use the ARIMA model to model the pressure data of the first 100ms, predict the pressure change trend and output 30% estimated compensation in advance, so as to control the steady-state pressure fluctuation within ±0.3N.

[0076] To prevent high-frequency electromagnetic interference (I_emi) from polluting the feedback signal, a dual suppression mechanism is adopted: Kalman filter denoising: When I_emi exceeds 30mV, the pressure feedback signal is subjected to noise suppression through the Kalman filter to improve the signal-to-noise ratio; Output attenuation coefficient: An attenuation coefficient (0.7-0.9) positively correlated with I_emi is applied to the PID output. The stronger the interference, the greater the attenuation, to avoid overcompensation caused by electromagnetic noise (such as misjudging low pressure and excessively increasing the feed rate).

[0077] After each feed adjustment (step length ≤ 0.002mm), a strict closed-loop verification process is performed: Continuously monitor the standard deviation of pressure fluctuations (σ_F) within 50ms. When σ_F ≤ 0.3N and is stable for three consecutive sampling periods (60ms), the compensation is considered complete. If the standard is not met, re-execute the parameter self-tuning process, with the cumulative number of iterations ≤ 5 times, to control the length of a single compensation cycle to meet the beat requirements of the automated production line.

[0078] Finally: The above description is only 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 in the scope of protection of the present invention.

Claims

1. A high-precision multi-interface coordinated FTU dynamic function test system, characterized in that: include: Protection unit: It is composed of an acrylic transparent protective cover and a sheet metal box. The acrylic transparent protective cover covers the top of the sheet metal box to form a test space. Test carrier unit: includes a quickly replaceable carrier plate, an access structure, and a positioning assembly. The access structure is driven by a rodless cylinder to achieve the in and out movement of the carrier plate on the slide rail. The positioning assembly includes a positioning pin and a guide block. Core test unit: includes a pressure plate structure, an endoscope, and a test module. The pressure plate structure is driven by a three-axis cylinder to achieve lifting and pressing actions. The endoscope and test module are located below the pressure plate structure. The test module integrates a button, short-circuit pin, network port, DB9, and aviation plug test module. The button test module is driven by a needle-type cylinder. Safety detection unit: including safety grating, anti-pinch device and pre-feeding sensing mechanism. The safety grating is set at the entrance of the test space; Control system: controls the start-up, execution and result uploading of the test process, and is equipped with PLC modules and database storage functions.

2. A high-precision multi-interface collaborative FTU dynamic function testing method, characterized in that: The system according to claim 1 comprises the following steps: Place the FTU product to be tested on the carrier, manually open the product lid, and trigger the pre-feed sensing mechanism to use a dual-wavelength collaborative detection mechanism for non-contact preliminary judgment, and then perform a secondary confirmation through physical contact to confirm that the lid is in the open state; Start the safety light barrier installed at the entrance of the test space, complete the optical path calibration, signal threshold configuration and linkage equipment communication test in sequence, and enter the detection state; The carrier plate is driven to move along the slide rail by a rodless cylinder. Positioning is performed by a positioning pin with a dual-stage floating structure integrated with a micro pressure sensor, and the positioning contact pressure is monitored in real time. Positioning deviations are automatically compensated by the fine-tuning screw of the guide block, and the carrier plate enters the test position. The three-axis cylinder drives the pressure plate structure downward, so that the buttons, short-circuit pins, network ports, DB9 and aviation plug test modules integrated in the pressure plate structure contact the interfaces of the FTU product to be tested in turn to perform performance testing; During the test, the endoscope collects test image data in real time, and the control system synchronously uploads the test data and image data to the host computer database and generates test results.

3. The method according to claim 2, characterized in that When the key test module performs performance testing, it includes: The material elastic coefficient and stroke threshold of the current test button are retrieved, and the driving air pressure of the needle cylinder is dynamically adjusted to the corresponding initial test pressure through the PID closed-loop control algorithm; The displacement S of the cylinder piston rod, the reaction force F of the key, and the response delay time T of the FTU during key triggering are collected in real time. Multi-dimensional correlation analysis is performed through the data fusion model to generate the key dynamic response curve. Perform nonlinear alignment between the key dynamic response curve and the standard curve, calculate the similarity score between the two, and determine the response performance; Among them, the extreme points of the first-order derivative of the key dynamic response curve are extracted to identify the risk of key sticking.

4. The method according to claim 3, characterized in that When generating a key dynamic response curve, it includes: The data fusion model receives normalized time series data of displacement S, reaction force F, and response delay time T through the input layer, extracts local features through a one-dimensional convolution with a convolution kernel size of 3×3 in the first hidden layer, captures long-term dependencies using a bidirectional LSTM in the second hidden layer, and dynamically allocates feature weights for the contact, trigger, and rebound test stages through the attention output layer. The final output is a key dynamic response curve including the trigger threshold point, peak force point, midpoint of the stable response segment, rebound starting point, and reset end point. Among them, during the model training stage, perturbations are added to the S, F, and T data through the fast gradient sign method to generate adversarial samples, and the dual loss function is used to optimize the model.

5. The method according to claim 2, characterized in that The short-circuit pin test module performs performance tests, including: The short-circuit pin test module uses a servo motor to drive a precision ball screw to achieve pin feed, collects the contact resistance value between the pin and the FTU interface in real time, and dynamically adjusts the feed speed based on a preset resistance threshold curve; When the contact resistance drops to the set value for the first time, the trigger pin pauses to eliminate contact bounce, and then continues to feed to the preset depth and maintains constant pressure to perform a short circuit test; If abnormal fluctuation occurs during the test, reverse retraction is started and the insertion process is re-executed; After the test is completed, the contact resistance time series data is feature extracted through the impedance spectrum analysis algorithm to generate the contact resistance fluctuation coefficient, which is then synchronously stored in the control system database.

6. The method according to claim 2, characterized in that When the network port test module performs performance testing, it includes: A servo motor is used to drive the network port test plug to feed along the Z axis. During the feeding process, the insertion force Fz and displacement Sz are collected in real time through the tension and pressure sensor to establish the Fz-Sz dynamic characteristic curve; When the insertion force Fz reaches the preset contact threshold F0, link negotiation of the first and second rate modes is performed in sequence, and the negotiation success rate and negotiation time are recorded; In the second rate mode, the built-in network analysis module injects a pseudo-random binary sequence to continuously monitor the bit error rate within a set time period and simultaneously collect the eye diagram parameters of the network port differential signal pair; If the bit error rate or eye diagram parameters do not meet the threshold requirements, the piezoelectric ceramic driver applies displacement compensation to the network port test plug, and the link negotiation and signal test are re-executed after each adjustment; After the test is completed, a comprehensive scoring report is generated and stored in the control system database.

7. The method according to claim 2, characterized in that The DB9 test module performs performance tests, including: The high-precision resistance measurement module uses a matrix scanning method to sequentially detect the contact resistance of each pin of the DB9 interface. Simultaneously, a high-precision differential probe is used to collect the transmission delay and signal rise time of the TX and RX signal pairs. When the contact resistance or signal rise time deviates from the set threshold, the piezoelectric ceramic driver applies displacement compensation to the DB9 interface and re-measures. During the electrical performance test phase, a simulated operating current is injected into the DB9 interface through a programmable load module to monitor the voltage drop and overcurrent protection response time in real time. Perform RS-232 and RS-485 protocol handshake tests in sequence, send data packets containing random check bits, record the protocol negotiation success rate and data transmission error rate, and generate a protocol compatibility matrix report; Among them, the collected multi-dimensional feature data is normalized, the anomaly score of each sample is calculated through the isolation forest model, abnormal samples are identified, and abnormal parameters are extracted.

8. The method according to claim 2, characterized in that When the aviation plug test module performs performance testing, it includes: The contact pressure, pin temperature and high-frequency interference current during the plugging process are collected synchronously through the micro pressure sensor, temperature sensor and high-frequency current probe integrated in the aviation plug test module; When the contact pressure fluctuates or the temperature exceeds the limit, the pressure compensation mechanism based on the fuzzy PID algorithm is activated to dynamically adjust the feed rate of the dual-axis servo motor, and the EMI filter is used to suppress high-frequency interference in real time.

9. The method according to claim 8, characterized in that Dynamically adjust the feed rate of the dual-axis servo motor, including: The contact pressure fluctuation, pin temperature change rate and high-frequency interference current amplitude are used as input variables of the fuzzy PID controller; A three-dimensional fuzzy rule base that supports online updating is established. The contact pressure fluctuation, pin temperature change rate, and high-frequency interference current amplitude are divided into multiple fuzzy subsets. The fuzzy output of the PID parameter correction is obtained through the centroid method. When the pressure fluctuation still exceeds the threshold after multiple consecutive compensations, the genetic algorithm is automatically triggered to optimize the membership function of the rule antecedent.

10. The method according to claim 9, characterized in that Dynamic adjustment of the feed rate of the dual-axis servo motor also includes: Dynamically switch the control mode according to the current pressure range: Incremental PID is used in the coarse adjustment stage, and position PID and feedforward compensation are enabled in the fine adjustment stage; When the high-frequency interference current amplitude exceeds the set value, the pressure feedback signal is noise-suppressed by the Kalman filter, and an attenuation coefficient is applied to the PID output; After each feed rate adjustment, the pressure fluctuation standard deviation is continuously monitored. When the pressure fluctuation standard deviation is stable below the set value for consecutive sampling periods, the compensation is considered complete. Otherwise, the parameter self-tuning process is re-executed.

Citation Information

Cited By

  • Test control method and system for direct-current power supply

    CN121633905A

  • GPS test equipment for controller, control method of GPS test equipment and manufacturing system of GPS test equipment

    CN121831820A