A test method for a hydraulic control check valve

Abstract

This invention relates to the field of hydraulic check valve testing technology and discloses a hydraulic check valve testing method. This method addresses the problem that existing testing methods, due to their limited scope of operation, insufficient data acquisition dimensions, and lack of dynamic modeling capabilities, fail to comprehensively reflect the true performance of hydraulic check valves under complex real-world conditions. By introducing a programmable operating condition loading module, a dynamic hydraulic testing environment is constructed based on pressure and flow rate change curves in actual applications. Combined with a high-frequency displacement sensor, a micro-flow sensor, and a vibration sensing unit, high-precision, multi-dimensional, synchronous data acquisition of the entire opening and closing process of the hydraulic check valve is achieved. This solution effectively compensates for the shortcomings of traditional testing methods in reproducing realistic dynamic operating conditions, significantly improving the authenticity of valve performance testing and the detailed completeness of response data. It provides strong support for accurately evaluating the dynamic behavior of hydraulic check valves under complex operating conditions.

Description

Technical Field

[0001] This invention relates to the field of hydraulic control check valve testing technology, specifically a hydraulic control check valve testing method. Background Technology

[0002] As a key component in hydraulic control systems, pilot-operated check valves are widely used in high-reliability applications such as engineering machinery, aerospace, and metallurgical equipment. They are primarily used to prevent reverse flow, maintain load stability, and achieve conditional reverse conduction under specific control signals. Their dynamic performance directly affects the system's response time, energy loss, and operational stability. Especially under complex conditions such as load-sensitive control and proportional synchronous control, pilot-operated check valves must maintain rapid and stable response characteristics under various non-constant pressure and flow conditions.

[0003] Currently, testing methods for hydraulic check valves are generally based on static or quasi-static operating conditions. These methods primarily involve setting constant pressure and fixed flow rates to measure basic performance parameters such as opening pressure, closing pressure, and leakage rate. These methods are simple to operate and use relatively simple equipment, making them widely applicable for valve factory inspection and routine performance evaluation. However, in actual hydraulic systems, valves often operate under significant dynamic characteristics, exhibiting complex boundary states such as sudden pressure increases, abrupt flow changes, high-frequency opening and closing, and superimposed system vibrations. Current testing methods struggle to effectively reproduce these typical operating conditions, leading to significant discrepancies between test results and actual performance.

[0004] The root cause of the above problems lies in the lack of the existing testing platform's ability to reproduce real dynamic operating conditions. This is mainly reflected in the following aspects: Limited operating condition loading methods: Traditional testing systems mostly adopt a quantitative pump plus overflow valve structure, which can only provide a constant pressure or slowly changing test environment, and cannot simulate transient behaviors such as sudden loading, periodic disturbances, and impact responses; Insufficient resolution and synchronization of the sensing system: Due to the lack of high-frequency, high-precision multi-dimensional sensing nodes, it is difficult to capture the dynamic characteristics of the valve core during rapid opening and closing, such as minute displacements, initial jitter, and vibration impacts, resulting in information loss and response delays in the test data; Lack of behavioral analysis mechanisms: Traditional evaluation mainly relies on the final static measurement value, and cannot establish a time-series behavioral model from control input to valve response, thus failing to identify potential risks such as action lag, incomplete opening and closing, and micro-leakage oscillations.

[0005] For the reasons mentioned above, even if a hydraulic check valve performs well in static testing, it may experience instability, self-excited oscillation, valve core jamming, or abnormal internal leakage under real-world operating conditions. In severe cases, this can lead to the entire hydraulic system becoming uncontrollable or failing. Especially in applications such as high-load holding, high-altitude work support, and high-frequency repetitive motion, such dynamic failures can pose safety hazards and economic losses. In summary, existing testing methods, due to their limited scope of operating conditions, insufficient data acquisition dimensions, and lack of dynamic modeling capabilities, cannot fully reflect the true performance of hydraulic check valves under complex real-world conditions. There is an urgent need for an advanced testing method that can cover multi-dimensional dynamic operating conditions, possess high-frequency behavior perception capabilities, and perform intelligent identification and trend analysis to improve testing accuracy and the reliability of engineering applications. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a testing method for hydraulic control check valves. By constructing a programmable dynamic operating condition loading environment and introducing a high-precision multi-dimensional sensing unit, it achieves accurate testing and comprehensive data acquisition of hydraulic control check valves under real dynamic conditions, thus solving the problems mentioned in the background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a test method for a hydraulically controlled check valve, comprising:

[0008] Select a pressure-time and flow-time curve that matches the application scenario of the hydraulic control check valve under test, and convert it into programmable control instructions through the working condition generation and loading module.

[0009] Connect the valve under test to a series test circuit and install a non-contact displacement sensor, a micro-flow sensor, and an acceleration sensor.

[0010] Start the servo hydraulic power source, apply pressure to the test circuit and control the flow according to the set working conditions, and collect the valve's opening and closing time, displacement trajectory and fluctuation amplitude during the response process under different pressure gradients and flow rates.

[0011] The collected data is modeled using a behavior analysis engine. By comparing behavior feature vectors, sluggish movements, incomplete opening and closing, abnormal leakage, and vibration characteristics are identified. The fault type and cause are inferred by combining historical models.

[0012] Output response curves, deviation analysis from standard models, risk assessment based on characteristic anomalies, and optimization suggestions for structural parameters and assembly accuracy;

[0013] Test data is archived into a behavioral feature database, and the model parameters are adaptively updated using a sliding window incremental learning mechanism.

[0014] Furthermore, the sources of the operating condition curves include:

[0015] By sensing and collecting data on the actual operation of hydraulic equipment in typical application scenarios, measured data of pressure-time and flow-time at different time periods are obtained.

[0016] Based on the modeling and simulation results of the hydraulic system, a dynamic response model is constructed according to the structural parameters and boundary conditions, and a virtual working condition with controllable characteristics is generated.

[0017] Call the preset standard working condition library, which is built according to common load modes in different industry applications and supports automatic retrieval and adaptation selection by working condition category;

[0018] The operating condition curve types include linear rise, sine wave, step response, sawtooth wave, random disturbance and multi-segment combination curve, which respectively simulate the dynamic response of the hydraulic control check valve under slow loading, periodic disturbance, sudden shock, unstable fluctuation and compound operating conditions.

[0019] Multi-segment composite curves construct complex operating conditions with loading, steady-state, impact, and unloading characteristics by switching different curve types according to time periods. All curves support parameter setting and splicing editing, serving as the input basis for generating programmable control instructions.

[0020] Furthermore, the test circuit adopts a series structure, with the hydraulically controlled check valve under test placed in the middle of the main circuit. The circuit includes an upstream pressure supply module consisting of an electro-hydraulic servo-driven pump and a proportional valve to stabilize the output pressure and flow rate; the bypass throttling device uses a proportional throttling valve to regulate the flow rate; the back pressure regulating unit is equipped with an adjustable back pressure valve to control the outlet pressure; and the pressure buffer assembly is equipped with an accumulator to suppress transient fluctuations.

[0021] The displacement sensor uses a laser triangulation type non-contact displacement sensor;

[0022] The micro-leakage flow sensor is installed in a separate channel at the outlet of the valve body's drain port, and uses a thermal flow sensor for measuring micro-flow rates.

[0023] The accelerometer is a triaxial miniature piezoelectric accelerometer;

[0024] All sensors are connected to a central data acquisition controller, and the timing synchronization of the sampling signals is achieved through a unified hardware clock and data caching mechanism.

[0025] Furthermore, the hydraulic power source adopts an electro-hydraulic servo-driven variable displacement piston pump, which, combined with a proportional relief valve and a proportional flow valve, forms a closed-loop pressure and flow control circuit.

[0026] Based on the programmed operating condition control curve, the hydraulic source and pressure regulating components are controlled by PLC or industrial PC to realize real-time adjustment of flow rate and pressure change gradient, simulating different load changes and unloading conditions.

[0027] The collected dynamic performance parameters include opening and closing delay time, valve core full-stroke displacement trajectory, pressure response amplitude, vibration and shock amplitude, micro-leakage change, and signal processing mechanism.

[0028] Furthermore, after normalizing the various response parameters collected in each test, a multidimensional feature vector with fixed dimensions is formed, and a weighting factor is introduced to weight the key indicators.

[0029] The valve core action response time deviates from the preset standard range, resulting in delayed opening and closing; the valve core displacement trajectory does not reach the full stroke threshold at the end, resulting in abnormal stroke termination; the oil drain port flow continuously exceeds the sealing failure threshold, resulting in abnormal leakage; the frequency domain energy density of the vibration signal is significantly higher than the benchmark model, indicating abnormal vibration characteristics.

[0030] By using the KNN nearest neighbor algorithm and cosine similarity evaluation mechanism, the current test behavior vector is compared with multiple labeled samples in the historical model library, and the anomaly category and its corresponding confidence score are output.

[0031] By combining test environment parameters, historical fault samples and current deviation patterns, inferences are generated, pointing out possible causes of abnormal behavior, including insufficient spring preload, valve core wear or contaminant jamming.

[0032] Furthermore, the behavior analysis engine constructs a multi-dimensional feature vector that includes multiple dimensions such as opening and closing time, maximum displacement, peak vibration acceleration, and leakage rate. After standardization, it is used to perform KNN comparison with historical samples. The similarity is measured by cosine angle, and the output is anomaly label and confidence score.

[0033] Furthermore, after the test is completed, the intelligent testing and analysis system generates a structured report containing response curves, performance evaluation, deviation analysis and improvement suggestions. The response curves include valve core displacement-time, pressure-time and leakage-time curves, which are derived from real-time data acquisition.

[0034] The standard model deviation analysis uses the least squares method to fit and calculate the difference in opening and closing time, displacement trajectory residual, leakage deviation ratio and vibration spectrum offset between the test data and the preset standard model.

[0035] Based on the weighted scoring model, thresholds and weights are set for each indicator, the overall performance score is evaluated and divided into three levels: high risk, medium risk and low risk, and processing suggestions are matched according to the scoring results.

[0036] The system combines behavioral characteristics with a fault tag library to automatically generate structural optimization suggestions, assembly accuracy correction schemes, and material matching suggestions.

[0037] Furthermore, a behavioral feature database is built. After each test, the system stores the test results in a structured record format. The record includes the timestamp information of the test time, the unique number of the specimen, the identification code of the applied loading condition, the original sensor response data including pressure, flow, displacement and vibration, and the behavioral feature vector obtained by processing based on preset rules.

[0038] The system also classifies and labels the corresponding anomaly identification results, including whether there is sluggish action, incomplete opening and closing, abnormal leakage or vibration, and assesses the risk level of the test sample according to the degree of deviation. Each record also includes the automatically inferred fault type label, the system recommended optimization suggestion number, as well as the corresponding test equipment number, operator identification and sampling batch number.

[0039] A sliding window incremental learning mechanism is introduced, in which the system continuously receives new test data, removes the oldest data, constructs an updated training set of the same size, and uses this to drive the parameter update of the lightweight supervised learning model.

[0040] Furthermore, the suggestion number is the processing suggestion ID automatically matched by the system based on the analysis results, which can be traced back to the specific measure entry in the knowledge base; the optimization tag is used to mark the structural parts that need to be optimized: springs, guide sleeves, and sealing rings, providing structural targets for closed-loop feedback.

[0041] Furthermore, the model employs support vector machines or Naive Bayes algorithms to dynamically optimize recognition capabilities by adjusting feature weights and classification boundaries.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. By introducing a programmable operating condition loading module, a hydraulic test environment with dynamic characteristics is constructed based on the pressure and flow change curves in actual applications. Combined with high-frequency displacement sensors, micro-flow sensors, and vibration sensing units, high-precision, multi-dimensional synchronous acquisition of the entire opening and closing process of the hydraulic control check valve is achieved. This solution effectively makes up for the shortcomings of traditional test methods in reproducing real dynamic operating conditions, significantly improves the authenticity of valve performance testing and the detail and completeness of response data, and provides strong support for accurately evaluating the dynamic behavior of hydraulic control check valves under complex operating conditions.

[0044] 2. By constructing a behavioral feature modeling and comparison mechanism, the response trajectory generated during the test can be intelligently analyzed to identify action lag, abnormal leakage and vibration characteristics. Combined with database archiving and incremental learning, it can realize fault trend prediction and performance degradation identification. At the same time, it supports structured report output and optimization suggestion generation, forming a closed-loop feedback path from testing to improvement, thereby improving the intelligence level and engineering adaptability of the hydraulic control check valve test system. Attached Figure Description

[0045] Figure 1 This is a flowchart of a test method for a hydraulically controlled check valve according to the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Please see Figure 1 A test method for a hydraulically controlled check valve, comprising:

[0048] Select a pressure-time and flow-time curve that matches the application scenario of the hydraulic control check valve under test, and convert it into programmable control instructions through the working condition generation and loading module.

[0049] Connect the valve under test to a series test circuit and install a non-contact displacement sensor, a micro-flow sensor, and an acceleration sensor.

[0050] Start the servo hydraulic power source, apply pressure to the test circuit and control the flow according to the set working conditions, and collect the valve's opening and closing time, displacement trajectory and fluctuation amplitude during the response process under different pressure gradients and flow rates.

[0051] The collected data is modeled using a behavior analysis engine. By comparing behavior feature vectors, sluggish movements, incomplete opening and closing, abnormal leakage, and vibration characteristics are identified. The fault type and cause are inferred by combining historical models.

[0052] Output response curves, deviation analysis from standard models, risk assessment based on characteristic anomalies, and optimization suggestions for structural parameters and assembly accuracy;

[0053] Test data is archived into a behavioral feature database, and the model parameters are adaptively updated using a sliding window incremental learning mechanism.

[0054] All sensors are connected to a central data acquisition controller, and the timing synchronization of the sampling signals is achieved through a unified hardware clock and data caching mechanism.

[0055] Select a pressure-time and flow-time curve that matches the application scenario of the hydraulically controlled check valve under test, and convert it into programmable control instructions through the operating condition generation and loading module. The specific implementation is as follows:

[0056] By sensing and collecting data on the actual operation of hydraulic equipment in typical application scenarios, measured pressure-time and flow-time data at different time periods are obtained. Based on the modeling and simulation results of the hydraulic system, a dynamic response model is constructed according to structural parameters and boundary conditions, and a virtual working condition with controllable characteristics is generated. A preset standard working condition library is called, which is established according to common load modes in different industry applications. It supports automatic retrieval and adaptation selection by working condition category. The working condition curves are mainly obtained from three clear channels. The first is the measured hydraulic equipment operation data. That is, by deploying high-precision pressure and flow sensors in typical applications of hydraulic control check valves, such as hydraulic systems of engineering machinery or industrial hydraulic stations, the actual data of pressure and flow changes over time in different operating stages such as start-up, operation, impact, and shutdown are collected and digitally recorded through data sampling cards to form the original working condition curve data.

[0057] The second category is simulation results based on mathematical models of hydraulic systems. This method typically uses fluid dynamics modeling tools (such as AMESim and SimHydraulics) to model and simulate typical operating processes of the system under test, based on known structural parameters, control logic, and boundary conditions. This yields dynamic response curves of the system under different operating conditions, which are then used to generate virtual operating conditions that approximate actual operating states. The third category is a pre-established database of standard operating condition curves. This database performs standardized analysis of common working cycles of hydraulic actuators according to different industry applications (such as agricultural equipment, lifting equipment, and high-speed injection molding systems), extracts representative typical pressure and flow rate change curves, and organizes them according to parameters such as application type, impact intensity, and frequency characteristics for use by the testing system.

[0058] The operating condition curve types include linear rising curves, sine waves, step response curves, sawtooth waves, random disturbance curves, and multi-segment combination curves. These simulate the dynamic response of the hydraulic check valve under slow loading, periodic disturbances, sudden shocks, unstable fluctuations, and combined operating conditions. To cover various fluid dynamic characteristics that the hydraulic check valve may encounter in actual operating conditions, the curve types used include: linear rising curves, used to simulate the gradual increase in pressure during slow loading of the hydraulic system, which can be used to detect the valve's opening and closing response threshold and opening / closing start point; and sine wave curves, used to simulate periodic pulsation or oscillation conditions, suitable for evaluating the valve's following behavior under frequent opening and closing. Performance and hysteresis effects; step response curves are used to reproduce the transient impact of a system under sudden loading or unloading in a short period of time, mainly used to test the instantaneous response capability and structural strength boundary of valves; sawtooth wave curves simulate periodic rise-sudden fall processes, often used to measure repositioning errors during opening and closing; random disturbance curves construct irregular pressure or flow disturbance sequences by setting random distributions of amplitude and frequency, simulating unpredictable fluctuations in operating conditions; multi-segment combination curves combine the above curves according to the time axis to form composite operating conditions, so as to more realistically reproduce the continuous and variable dynamic state in complex workflows. Each type of curve needs to set clear parameter indicators, such as time constant (i.e., the proportion of time required to reach a stable value), slope of change (pressure or flow increase per unit time), fluctuation frequency (number of cycles of sinusoidal or periodic changes), and peak amplitude (maximum pressure or flow value). These parameters determine the valve's response performance under different operating conditions.

[0059] The multi-segment composite curves construct complex operating conditions with loading, steady-state, shock, and unloading characteristics by switching different curve types according to time periods. All curves support parameter setting and splicing editing, serving as input for generating programmable control instructions. The loading segment simulates the initial pressurization or power-on loading process of the system, typically accompanied by a gradual increase in pressure or flow, suitable for observing the initial action point of the valve core. The steady-state segment maintains a continuous and stable pressure and flow, examining the valve's internal leakage behavior and structural stability under continuous operation. The shock segment introduces steep pressure changes or high-frequency fluctuations to test the valve's structural integrity and anti-interference response under dynamic disturbance conditions. The unloading segment simulates scenarios of rapid pressure release or sudden flow reduction, evaluating the valve's return capability and hysteresis characteristics. Each segment can be set independently via the system, including the rate of change (the magnitude of change per unit time), peak limit (maximum output pressure / flow), and response period (duration of maintaining the segment). User-defined editing and sequence adjustment are also supported to create various dynamic test conditions.

[0060] The test loop adopts a series structure, with the hydraulically controlled check valve under test placed in the middle of the main loop. The loop includes an upstream pressure supply module consisting of an electro-hydraulic servo-driven pump and a proportional valve to stabilize the output pressure and flow rate. A bypass throttling device uses a proportional throttling valve to regulate the flow rate. The back pressure regulating unit is equipped with an adjustable back pressure valve to control the outlet pressure. The pressure buffer component is equipped with an accumulator to suppress transient fluctuations. Specifically, the system discretely samples the curve data at fixed time intervals to form a series of control point sequences, each point corresponding to a target pressure or flow rate value at a given time point. The sampling period is typically set between 10ms and 50ms and can be adjusted according to the response speed of the valve under test. Subsequently, the system converts this control point sequence into corresponding analog control commands in real time, specifically the voltage signal (0–10V) or current signal (4–20mA) required by the servo drive system, and sends them to the pump drive unit or proportional valve actuator through a closed-loop controller. During control execution, the system also adjusts the deviation based on the real-time feedback pressure / flow rate values, thereby achieving precise and stable dynamic hydraulic condition application and ensuring that the simulation output closely matches the target curve setting.

[0061] The valve under test is connected to a series test circuit, and a non-contact displacement sensor, a micro-flow sensor, and an acceleration sensor are installed. The specific implementation is as follows:

[0062] The test circuit adopts a series hydraulic test circuit structure, in which the hydraulically controlled check valve under test is located in the middle of the main circuit. The test circuit also includes an upstream pressure supply module, a bypass throttling device, a back pressure adjustment unit, and a pressure buffer assembly to ensure a constant pressure source and adjustable back pressure during testing. The hydraulic circuit design of the test system uses a series arrangement, meaning that the hydraulically controlled check valve under test is directly connected in series as a core component of the main test circuit. It is not placed in a branch or parallel module; this arrangement more closely resembles its position in a real hydraulic system. To ensure stable and controllable operating conditions, the entire test circuit is equipped with an upstream pressure supply module, typically composed of a variable pump with servo regulation and a proportional solenoid valve, which can continuously provide stable pressure. A bypass throttling device is also included to simulate the throttling effect caused by the circuit characteristics in an actual system. This device can adjust its opening degree through an electro-proportional controller, thereby precisely controlling the flow rate. The back pressure adjustment unit is mainly used to adjust the back pressure value on the valve outlet side, using a back pressure valve with electro-pneumatic feedback to make the outlet pressure difference controllable, suitable for simulating different action processes. In addition, a pressure buffer component, such as an accumulator, is installed between the pump outlet and the valve inlet. Its main function is to reduce pressure fluctuations, prevent transient peaks from interfering with sensor data, and ensure that the entire test loop can maintain stable input and output boundary conditions during dynamic testing.

[0063] The displacement sensor uses a laser triangulation-type non-contact displacement sensor, installed on the outside of the transparent valve cavity observation window. It provides high-precision tracking of the axial movement of the valve core. The opening and closing process of the valve core is one of the core performance indicators in the evaluation of hydraulic check valves; therefore, a laser triangulation-type non-contact displacement sensor was chosen for tracking. This type of sensor is widely used in industrial measurement due to its high sensitivity and high repeatability, and is particularly suitable for applications with limited structural space or where contact with the measurement surface is not possible. The installation position is chosen on the outside of the transparent observation window on the valve body. This window is made of high-strength, pressure-resistant acrylic or tempered glass and has undergone anti-fogging and anti-reflection treatment, which can withstand hydraulic shocks and ensure stable transmission of the laser beam under high-frequency vibrations. A safe distance is maintained between the sensor body and the valve cavity, and a protective cover is installed to prevent damage from oil mist or hydraulic shocks. The sensor has a sampling frequency of at least 5kHz, which means it can acquire 5,000 position data points per second. With a resolution better than 5 micrometers, it can clearly reproduce the minute movement trajectory of the valve core from the starting point to full opening and then to closing. It is especially suitable for identifying opening and closing jams, hysteresis displacements, or staged jump phenomena.

[0064] The micro-leakage flow sensor is installed in a separate channel at the outlet of the valve body's drain port, and uses a thermal flow sensor for measuring micro-flow rates.

[0065] A minute amount of internal leakage may occur during the operation of a hydraulic check valve. This is a crucial parameter for assessing its sealing condition, guide structure matching, and manufacturing precision. Therefore, a dedicated micro-flow measurement channel is installed at the valve body's outlet. This channel is isolated from the main circuit to ensure it is not affected by main flow pulsations. The channel is straight and short to minimize additional pressure drop. A thermal micro-flow sensor is selected. This type of sensor detects fluid flow rate by relying on the temperature difference between the heating element and the thermocouple. It has extremely high sensitivity and is particularly suitable for detecting low-speed, low-volume flow of hydraulic oil. It can measure a minimum flow rate of less than 0.01 liters per minute, meaning that even slight sealing defects or the initial stage of internal leakage can be detected promptly. Its response time is less than 10 milliseconds, meaning it can capture minute leakage changes in a short period after the valve core closes in real time. This configuration greatly enhances the ability to identify latent internal leakage failures, preventing unforeseen oil consumption, energy consumption, or secondary malfunctions during later use.

[0066] The accelerometer is a triaxial miniature piezoelectric accelerometer, fixed to the metal structure on top of the valve body. It monitors the impact vibration characteristics generated during the opening and closing of the valve core. The valve core exhibits not only displacement characteristics during opening and closing but also generates minute impacts and high-frequency vibrations at the opening / closing boundaries or when excited by oil flow. These signals are crucial for identifying motion quality, structural resonance problems, jamming, or rebound phenomena. Therefore, a triaxial piezoelectric accelerometer is fixedly installed on the metal housing on top of the valve body. This sensor is small and lightweight, with minimal impact on the overall stiffness of the valve body after installation, yet it can highly sensitively detect acceleration signals from the X, Y, and Z directions. The fixing method uses a dedicated threaded seat or adhesive base to ensure a stable and reliable vibration signal transmission path. The sensor has a sampling frequency of no less than 10kHz, which can fully cover the typical vibration range of hydraulic control systems from 1Hz to several kilohertz. In particular, it responds very quickly to impact acceleration and can be used to extract the impact intensity, resonance fluctuations and dynamic stiffness changes during the opening and closing process, thereby supplementing the dynamic transient response behavior that displacement sensors cannot observe.

[0067] All sensors are connected to a central data acquisition controller. A unified hardware clock and data caching mechanism ensure time synchronization of the sampled signals. To guarantee time alignment of signals from sensors across different physical dimensions, a central data acquisition controller serves as the unified data entry point. Each sensor's output channel connects to this controller via a high-speed interface. The controller incorporates a hardware-level unified clock source and a data buffer unit. All sampled data is timestamped before entering the main buffer. This mechanism guarantees precise time-axis correspondence between different signal streams, with a maximum error of less than 1 millisecond. This ensures that subsequent data processing and modeling analysis can reconstruct the true causal relationship between valve core movement, leakage changes, and vibration impacts along the timeline. This unified acquisition architecture is also a fundamental condition for subsequent intelligent behavior modeling and multi-dimensional fault identification.

[0068] The servo hydraulic power source is activated, and pressure is applied to the test circuit and the flow rate is controlled according to the set operating conditions. Performance parameters such as valve opening and closing time, displacement trajectory, and fluctuation amplitude during the response process are collected under different pressure gradients and flow rates. Specifically, the implementation is as follows:

[0069] The hydraulic power source employs an electro-hydraulic servo-driven variable displacement piston pump, combined with a proportional relief valve and a proportional flow valve to form a closed-loop pressure and flow control circuit. By activating the high-performance electro-hydraulic servo hydraulic power source, precisely controlled pressure and flow loads are dynamically applied to the test circuit according to pre-set hydraulic conditions to simulate the complex dynamic response of a hydraulically controlled check valve under actual operating conditions. The hydraulic power source used is an electro-hydraulic servo-driven variable displacement piston pump, which can continuously adjust the output displacement according to the control signal. Combined with the proportional relief valve and proportional flow valve, it forms a dual closed-loop control system, achieving high-precision real-time adjustment of circuit pressure and flow. The system control accuracy is better than ±0.5% of full scale and has a fast response capability, with a dynamic response time of less than 50 milliseconds, ensuring that it can quickly and accurately follow changes in the set operating conditions and achieve precise application of different pressure gradients and flow rates.

[0070] Based on the programmed operating condition control curve, the hydraulic source and pressure regulating components are controlled by a PLC or industrial PC to achieve real-time adjustment of flow rate and pressure change gradient, simulating different load changes and unloading conditions. During the load application process, the PLC or industrial computer executes the programmed pressure-time and flow-time control curves to adjust the hydraulic source and regulating valve in real time, simulating common load changes, unloading and complex waveform conditions in the field, covering a variety of typical operating conditions such as sudden pressure increase, slow unloading, and periodic fluctuations, greatly improving the authenticity and representativeness of the test.

[0071] The collected dynamic performance parameters include opening / closing delay time, valve core displacement trajectory throughout its entire stroke, pressure response amplitude, vibration and shock amplitude, minute leakage change, and signal processing mechanism. During testing, the system simultaneously collects multiple dynamic performance indicators of the valve, including opening / closing delay time (the time from control signal triggering to initial valve core movement), valve core displacement trajectory throughout its entire stroke (displacement-time curve obtained through displacement sensors, used to calculate velocity, acceleration, and identify jamming phenomena), pressure response amplitude (the pressure difference change measured before and after valve action), vibration and shock amplitude (instantaneous peak value and its frequency domain energy recorded by an acceleration sensor), and minute leakage (the accumulated leakage flow within 1 second after valve closure). These parameters comprehensively reflect the valve's performance under actual dynamic conditions, providing a scientific basis for valve condition analysis, fault diagnosis, and optimized design.

[0072] Signal processing mechanism: After the raw sampled data undergoes first-order low-pass filtering, moving average processing, waveform envelope extraction, and spectral transformation analysis, feature values ​​are extracted for subsequent modeling. To improve the accuracy and usability of the data, the acquired raw signal is first subjected to first-order low-pass filtering to remove high-frequency noise, then the data is smoothed using a moving average algorithm, followed by waveform envelope extraction to obtain the signal's envelope features, and finally, spectral transformation is used to perform frequency domain analysis on the signal to extract the frequency components in the vibration and pressure signals. These processed feature parameters will serve as the basis for subsequent behavioral modeling and intelligent analysis, enabling in-depth evaluation of the performance of the hydraulic check valve.

[0073] The collected data is modeled using a behavioral analysis engine. By comparing behavioral feature vectors, characteristics such as sluggish movements, incomplete opening and closing, abnormal leakage, and vibration are identified. Combined with historical models, the fault type and cause are inferred. Specifically, the implementation is as follows:

[0074] After normalizing the various response parameters collected in each test, a multidimensional feature vector with fixed dimensions is formed. Weighting factors are then introduced to weight key indicators. To achieve intelligent analysis and anomaly identification of hydraulic check valve test data, the system employs an advanced behavioral analysis engine to perform multidimensional modeling of the collected dynamic performance data, forming a behavioral feature vector. First, multiple key response parameters acquired during the test, including opening and closing time, vibration frequency, and leakage rate, are normalized and converted into indicators with the same numerical scale to ensure the comparability of data with different dimensions. Subsequently, based on the importance of each parameter to valve performance, weighting factors are introduced for weighted calculation, giving key indicators a greater proportion in the feature vector and improving the accuracy and sensitivity of anomaly identification.

[0075] The valve core's response time deviates from the preset standard range, resulting in delayed opening and closing; the valve core's displacement trajectory fails to reach the full stroke threshold at the end, indicating abnormal stroke termination; the drain port flow rate continuously exceeds the sealing failure threshold, indicating abnormal leakage; the vibration signal's frequency domain energy density is significantly higher than the benchmark model, indicating abnormal vibration characteristics. For the identified abnormal types, the system sets strict discrimination thresholds to distinguish between normal and abnormal states. Incomplete opening and closing is manifested as the displacement trajectory not reaching more than 95% of the full stroke at the end; a static leakage rate greater than 0.05 liters per minute is considered abnormal internal leakage; and a high-frequency vibration signal energy density exceeding 1.5 times the normal benchmark value is considered abnormal vibration. These criteria are determined based on extensive experimental data and field experience, ensuring a high degree of identification capability and practicality for common fault types.

[0076] By employing the K-Nearest Neighbor (KNN) algorithm and cosine similarity evaluation mechanism, the current test behavior vector is compared with multiple labeled samples in the historical model database. The system outputs the anomaly category and its corresponding confidence score. The behavior analysis engine constructs a multi-dimensional feature vector including opening / closing time, maximum displacement, peak vibration acceleration, and leakage rate. After standardization, this vector is used for KNN comparison with historical samples. The similarity is measured using a cosine similarity angle, outputting anomaly labels and confidence scores. In terms of model comparison, the system uses the KNN algorithm to perform similarity analysis on the current test behavior feature vector, combined with a cosine similarity index to evaluate the similarity between the test vector and known fault label samples in the historical database, accurately locating the anomaly category corresponding to the current test behavior. Simultaneously, the system outputs a confidence score for the anomaly determination, quantifying the reliability of the identification results and assisting technicians in making more accurate maintenance decisions.

[0077] By combining test environment parameters, historical fault samples, and current deviation patterns, the system generates inference conclusions, indicating possible causes of abnormal behavior, including insufficient spring preload, valve core wear, or contaminant buildup. Finally, based on test environment parameters and a rich database of historical fault samples, the system automatically infers possible fault causes. For example, if sluggish action is detected, the system may infer that insufficient spring preload is causing a delayed valve core response; incomplete opening and closing may be due to valve core wear or contaminant buildup; abnormal internal leakage may be caused by damaged sealing surfaces or improper assembly. This intelligent inference mechanism helps users accurately pinpoint the root cause of problems, guiding subsequent maintenance and structural optimization, effectively improving the overall reliability and service life of the valve.

[0078] The output response curve, deviation analysis from the standard model, risk assessment based on characteristic anomalies, and optimization suggestions for structural parameters and assembly accuracy are implemented as follows:

[0079] After testing, the intelligent testing and analysis system generates a structured report containing response curves, performance evaluation, deviation analysis, and improvement suggestions. The response curves include valve core displacement-time, pressure-time, and leakage-time curves, derived from real-time data acquisition. The test report is divided into six independent and logically clear parts: First, the test information section records basic information such as test time, equipment number, operators, and test environment; second, the input conditions section specifies the parameters and variation characteristics of the pressure-time and flow-time curves applied during the test; third, the response curves display key curve graphs such as valve opening and closing time, valve core displacement trajectory, and pressure fluctuations during the test; fourth, the summary of valve performance index data, covering key quantitative parameters such as opening and closing delay, leakage, and vibration amplitude; fifth, deviation analysis, which calculates the deviation amplitude of opening and closing time, residual value of displacement trajectory, proportion of leakage exceeding the standard, and abnormal characteristics in the vibration spectrum against the standard model, comprehensively reflecting the deviation of valve performance; and finally, the fault conclusion and risk assessment, helping technicians quickly locate problems and determine the risk level.

[0080] The standard model deviation analysis uses the least squares method for fitting, calculating the difference in opening and closing time, displacement trajectory residual, leakage deviation ratio, and vibration spectrum offset between the test data and the preset standard model. Each index in the deviation analysis is quantitatively measured through a clear calculation method. The opening and closing time deviation refers to the difference between the test value and the standard action time. The displacement trajectory residual expresses the degree of deviation between the actual movement path of the valve core and the ideal path through the curve fitting error. The leakage excess ratio is the percentage of tests in which the valve leakage exceeds the set safety threshold after closing. The degree of vibration spectrum abnormality is calculated based on the peak amplitude and energy distribution in the frequency domain analysis results.

[0081] Based on a weighted scoring model, thresholds and weights are set for each indicator, the overall performance score is evaluated and divided into three levels: high risk, medium risk, and low risk. Based on the scoring results, processing suggestions are matched, and the threshold ranges and recommended processing strategies corresponding to each risk level are defined. Based on these indicators, the system uses a multi-indicator weighted scoring model to classify the comprehensive performance of the hydraulic check valve into three risk levels: high risk, medium risk, and low risk. Strict threshold ranges are defined for each level to ensure that the risk classification is scientific and reasonable, and corresponding processing strategies are recommended for users' reference.

[0082] The system combines behavioral characteristics with a fault tag library to automatically generate structural optimization suggestions, assembly accuracy correction schemes, and material matching suggestions. The generation of intelligent optimization suggestions relies on the system's built-in professional knowledge base and behavioral analysis results based on test data. It automatically identifies potential problems in the valve's structure and assembly and proposes specific improvement measures. For example, for the problem of insufficient spring stiffness causing slow opening and closing response, the system will suggest adjusting or replacing the spring to improve the elastic coefficient; if excessive guide clearance causes jamming, it will recommend modifying the guide structure to ensure smooth valve core movement; for seal coaxiality deviation, the system will indicate the direction of adjustment in the assembly process; for performance degradation caused by insufficient material wear resistance, it suggests optimizing the coating material or processing technology. The entire suggestion output process considers both the valve's physical structural characteristics and dynamic performance, achieving a closed-loop feedback from test data to design improvement, greatly improving the reliability and performance stability of subsequent products.

[0083] Test data is archived into a behavioral feature database, and the model parameters are adaptively updated using a sliding window incremental learning mechanism. The specific implementation is as follows:

[0084] The system stores the test results in a structured record format after each test. The record includes the timestamp of the test, the unique number of the specimen, the applied load condition identifier, the raw sensor response data (including pressure, flow rate, displacement, and vibration), and the behavioral feature vector obtained based on preset rules. The system also categorizes and labels the corresponding anomaly identification results, including whether there is sluggish action, incomplete opening / closing, abnormal leakage, or vibration. It assesses the risk level of the test sample based on the degree of deviation. Each record also includes an automatically inferred fault type label, a system-recommended optimization suggestion number, and the corresponding test equipment number, operator identification, and sampling batch number. To achieve intelligent evaluation and trend prediction of the dynamic response behavior of the hydraulic check valve, this system has constructed a complete data archiving and continuous model learning mechanism. All raw response data and analysis results acquired during the test are synchronously archived in the hydraulic check valve behavioral feature database to support subsequent analysis, modeling, and knowledge accumulation.

[0085] The suggested ID is a processing suggestion automatically matched by the system based on the analysis results, traceable to specific measure entries in the knowledge base; the optimization tag is used to mark the structural parts that need optimization: springs, guide sleeves, and sealing rings, providing structural targets for closed-loop feedback. This database design adopts a structured field hierarchical recording method. Each test record contains 13 key fields: test time (accurate to the second), specimen number (corresponding to the product's unique identifier), loading condition code (corresponding to the specific pressure-flow curve number), raw response data (including raw sensor sampling data such as displacement, pressure, vibration, and flow), and processed feature vector (data after filtering and extraction). The system includes fields for: digital behavioral characteristics, anomaly identification results (indicators of issues such as delayed opening / closing or abnormal leakage), risk level (a 1-5 level classification based on the degree of behavioral deviation), fault type label (fault type identified based on the association between abnormal characteristics and behavioral models, such as excessive internal leakage, abnormal jamming, poor sealing, etc.), recommended action number (matching the recommended repair or adjustment plan), and data source identifier (recording the data acquisition equipment, operator information, and test batch number). Other fields include supplementary items such as test conclusions, environmental conditions, and evaluation notes. All fields support combined retrieval and conditional filtering, facilitating multidimensional statistics and sample reuse in the later stages.

[0086] By introducing a sliding window incremental learning mechanism, the system continuously receives new test data, removes the oldest data, and constructs an updated training set of the same size. This updated training set drives the parameter updates of a lightweight supervised learning model. To improve the model's adaptability to complex operating conditions and diverse valve types, this system integrates an incremental learning method based on the sliding window mechanism. Every 20 new test records received, a model update process is automatically triggered. During the update, the system calls a lightweight supervised learning model (such as Naive Bayes, Support Vector Machine, or shallow neural network), using feature-label pairs from the new data as samples to fine-tune the model's classification threshold, loss function weights, and parameter matrix. This allows the model to gradually develop an understanding of the behavioral characteristics of new products, enabling real-time tracking and optimization, and avoiding recognition bias caused by overfitting to old samples.

[0087] This model employs Support Vector Machine (SVM) or Naive Bayes algorithms to dynamically optimize recognition capabilities by adjusting feature weights and classification boundaries. The core classifier uses SVM or Naive Bayes to identify the behavioral characteristics and fault modes of hydraulic check valves. The SVM model acts like drawing a reasonable "boundary" in the feature space, clearly separating the valve's normal state from various abnormal states (such as slow opening / closing, excessive oil leakage, and excessive vibration). The system determines whether a new test sample falls on the "boundary line" and is considered normal or abnormal. It can also handle non-linear feature relationships through kernel functions, accurately identifying even complex features with many dimensions. The Naive Bayes algorithm, on the other hand, assumes that each feature is relatively independent. By calculating the probability of each behavior (normal or a certain type of abnormality) in the features, it determines the likelihood of a new sample belonging to a particular category. This classification method is fast and computationally efficient, making it suitable for scenarios with many samples and moderate feature dimensions. During incremental learning, the system adaptively updates the weight parameters of key features (such as start-stop delay, leakage rate, vibration frequency, etc.) based on each round of newly received samples. At the same time, it fine-tunes the classification boundary of SVM or the prior probability of Naive Bayes, thereby dynamically improving the model's ability to identify new abnormal samples or samples with ambiguous boundaries. This ensures that the model continuously adapts to test data, accurately judges, and maintains the stability and robustness of the recognition results.

[0088] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0089] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0090] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0091] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0092] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0093] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0094] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0095] In conclusion, 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 within the protection scope of the present invention.

Claims

1. A test method for a hydraulically controlled check valve, characterized in that, include: Select a pressure-time and flow-time curve that matches the application scenario of the hydraulic control check valve under test, and convert it into programmable control instructions through the working condition generation and loading module. Connect the hydraulic control check valve under test to a series test circuit and install a non-contact displacement sensor, a micro-leakage flow sensor and an acceleration sensor. Start the servo hydraulic power source, apply pressure to the test circuit and control the flow according to the set working conditions, and collect the performance parameters of the test hydraulic control check valve under different pressure gradients and flow rates, such as opening and closing time, displacement trajectory and fluctuation amplitude during the response process. The collected data is modeled using a behavior analysis engine. By comparing behavior feature vectors, sluggish movements, incomplete opening and closing, abnormal leakage, and vibration characteristics are identified. The fault type and cause are inferred by combining historical models. Output response curves, deviation analysis from standard models, risk assessment based on characteristic anomalies, and optimization suggestions for structural parameters and assembly accuracy; Test data is archived into a behavioral feature database, and a sliding window incremental learning mechanism is used to adaptively update the parameters of the lightweight supervised learning model.

2. The test method for a hydraulically controlled check valve according to claim 1, characterized in that: The sources of pressure-time and flow-time curves include: By sensing and collecting data on the actual operation of hydraulic equipment in engineering machinery hydraulic systems and industrial hydraulic stations, measured data of pressure-time and flow-time at different time periods are obtained. Based on the modeling and simulation results of the hydraulic system, a dynamic response model is constructed according to the structural parameters and boundary conditions, and a virtual working condition is generated. Call the preset standard operating condition library, which supports automatic retrieval and adaptation selection by operating condition category; The operating condition curve types include linear rise, sine wave, step response, sawtooth wave, and random disturbance curve, which respectively simulate the dynamic response of the hydraulic check valve under slow loading, periodic disturbance, sudden shock, unstable fluctuation and compound operating conditions. Multi-segment composite curves construct complex operating conditions with loading, steady-state, impact, and unloading characteristics by switching different curve types according to time periods. All curves support parameter setting and splicing editing, serving as the input basis for generating programmable control instructions.

3. The test method for a hydraulically controlled check valve according to claim 1, characterized in that: The test circuit adopts a series structure, with the hydraulically controlled check valve under test placed in the middle of the main circuit. The test circuit includes an upstream pressure supply module consisting of an electro-hydraulic servo-driven pump and a proportional valve to stabilize the output pressure and flow rate; the bypass throttling device uses a proportional throttling valve to regulate the flow rate; the back pressure regulating unit is equipped with an adjustable back pressure valve to control the outlet pressure; and the pressure buffer assembly is equipped with an accumulator to suppress transient fluctuations. The displacement sensor uses a laser triangulation type non-contact displacement sensor; The micro-leakage flow sensor is installed in a separate channel at the outlet of the oil drain port of the hydraulic control check valve body being measured, and a thermal flow sensor for measuring micro-flow is used. The accelerometer is a triaxial miniature piezoelectric accelerometer; All sensors are connected to a central data acquisition controller, and the timing synchronization of the sampling signals is achieved through a unified hardware clock and data caching mechanism.

4. The test method for a hydraulically controlled check valve according to claim 1, characterized in that: The hydraulic power source uses an electro-hydraulic servo-driven variable displacement piston pump, which, together with a proportional relief valve and a proportional flow valve, forms a closed-loop pressure and flow control circuit. Based on the programmed operating condition control curve, the hydraulic source and pressure regulating components are controlled by PLC or industrial PC to adjust the flow rate and pressure change gradient in real time, simulating different load changes and unloading conditions. The collected dynamic performance parameters include opening and closing delay time, displacement trajectory of the valve core of the tested hydraulic control check valve throughout its full stroke, pressure response amplitude, vibration and shock amplitude, and changes in trace leakage.

5. The test method for a hydraulically controlled check valve according to claim 1, characterized in that: After normalizing the various response parameters collected in each test, a behavior feature vector with fixed dimensions is formed, and a weighting factor is introduced to weight the key indicators. The valve core action response time deviates from the preset standard range, resulting in delayed opening and closing; the valve core displacement trajectory does not reach the full stroke threshold at the end, resulting in abnormal stroke termination; the oil drain port flow continuously exceeds the sealing failure threshold, resulting in abnormal leakage; the frequency domain energy density of the vibration signal is higher than the benchmark model, indicating abnormal vibration characteristics. By using the KNN nearest neighbor algorithm and cosine similarity evaluation mechanism, the current test behavior vector is compared with multiple labeled samples in the historical model library, and the anomaly category and its corresponding confidence score are output. By combining test environment parameters, historical fault samples and current deviation patterns, inferences are generated to indicate the causes of abnormal behavior, including insufficient spring preload, valve core wear or contaminant jamming.

6. The test method for a hydraulically controlled check valve according to claim 1, characterized in that: The behavioral feature vector constructed by the behavioral analysis engine includes multiple dimensions such as opening and closing time, maximum displacement, peak vibration acceleration, and leakage rate. After standardization, it is used to compare with historical samples using KNN. The similarity is measured by cosine angle, and the output includes anomaly labels and confidence scores.

7. The test method for a hydraulically controlled check valve according to claim 1, characterized in that: After the test is completed, the intelligent testing and analysis system generates a structured report containing response curves, performance evaluation, deviation analysis and improvement suggestions. The response curves include valve core displacement-time, pressure-time and leakage-time curves, which are derived from real-time data acquisition. The standard model deviation analysis uses the least squares method to fit and calculate the difference in opening and closing time, displacement trajectory residual, leakage deviation ratio and vibration spectrum offset between the test data and the preset standard model. Based on the weighted scoring model, a threshold and weight are set for each indicator, the overall performance score is evaluated and divided into three levels: high risk, medium risk and low risk, and processing suggestions are matched according to the scoring results. The system combines behavioral characteristics with a fault tag library to automatically generate structural optimization suggestions, assembly accuracy correction schemes, and material matching suggestions.

8. The test method for a hydraulically controlled check valve according to claim 1, characterized in that: A behavioral feature database is built. After each test, the system stores the test results in a structured record format. The record includes the timestamp information of the test time, the unique number of the specimen, the identification code of the applied loading condition, the original sensor response data including pressure, flow, displacement and vibration, and the behavioral feature vector obtained by processing based on preset rules. The system also classifies and labels the corresponding anomaly identification results, including whether there are slow actions, incomplete opening and closing, abnormal leaks or vibrations, and assesses the risk level of the test sample according to the degree of deviation. Each record also includes the automatically inferred fault type label, the system recommended optimization suggestion number, as well as the corresponding test equipment number, operator identification and sampling batch number. A sliding window incremental learning mechanism is introduced, in which the system continuously receives new test data, removes the oldest data, constructs an updated training set of the same size, and uses this to drive the parameter update of the lightweight supervised learning model.

9. A test method for a hydraulically controlled check valve according to claim 8, characterized in that: The suggestion number is a processing suggestion ID automatically matched by the system based on the analysis results, which can be traced back to the specific measure entry in the knowledge base.

10. A test method for a hydraulically controlled check valve according to claim 8, characterized in that: Lightweight supervised learning models employ support vector machines or Naive Bayes algorithms to dynamically optimize recognition capabilities by adjusting feature weights and classification boundaries.

Images