Self-calibration and self-adaptive learning intelligent corrugated pipe test system and method

Abstract

The invention discloses a self-calibration and self-adaptive learning intelligent corrugated pipe test system and method, and belongs to the technical field of test devices. The system comprises an intelligent control module, a benchmark reference module, a data acquisition module and a basic fluid circulation module, online self-calibration is realized through the intelligent control module, and the benchmark reference module is used for comparing a theoretical value with a measured value to verify the instrument precision; an environment dynamic compensation function is introduced, and a heat loss model and fluid physical property parameters are corrected in real time; the adaptive learning mechanism generates a real-time compensation coefficient through historical data analysis, and system parameters are optimized; the method comprises the steps of self-calibration, accurate testing and adaptive learning, and full-process intelligent testing is achieved. According to the invention, full-process intelligentization from online calibration, accurate test to adaptive optimization is realized, and the accuracy and long-term stability of a test result are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of testing device technology, and more specifically, relates to a self-calibrating and adaptive learning intelligent bellows testing system and method. Background Technology

[0002] As an important flexible connection and thermal compensation element, the heat transfer performance and flow resistance characteristics of bellows are core indicators for evaluating product quality. Their performance directly affects the safety, efficiency, and reliability of the entire system. Therefore, it is crucial to accurately test the thermal conductivity and flow resistance characteristics of bellows before they leave the factory.

[0003] In existing technologies, bellows testing devices pump a constant-temperature, constant-pressure fluid (such as water) into the bellows under test through an inlet input mechanism, and use temperature and pressure sensors at the inlet and outlet ends to measure relevant data, calculating the temperature difference (…). ) and pressure drop ( The performance of such testing devices is evaluated using static measurement models. However, these models have several technical shortcomings that urgently need to be addressed in industrial applications and R&D scenarios that demand high precision and reliability: The core accuracy of the testing equipment relies on the long-term stability of temperature and pressure sensors. Time-varying errors in sensors, such as zero-point drift and temperature drift, cannot be detected and calibrated within the testing system, especially under conditions of drastic temperature changes. Existing equipment typically lacks effective online self-calibration capabilities, relying solely on periodic, manually operated external verification to ensure accuracy. This not only compromises the reliability of test data within the verification cycle but also affects the continuous use of the equipment. For R&D or quality control scenarios requiring high-precision testing, the inability to verify instrument status in real time is a significant weakness.

[0004] Existing equipment testing models (such as flow resistance and heat transfer calculations) are typically based on fixed parameters under ideal laboratory conditions, lacking real-time sensing and compensation mechanisms for changes in ambient temperature, humidity, and atmospheric pressure. However, fluid properties (density, viscosity) and system heat loss are closely related to environmental conditions, leading to systematic deviations in test results for the same product in different seasons and regions. The repeatability and accuracy of test data are difficult to guarantee, greatly affecting the uniformity of testing standards.

[0005] The core algorithms and model parameters of existing devices are fixed after leaving the factory. Their function is focused on a single "measurement-output" process. It is a static system, and its accuracy will irreversibly decrease as the equipment ages. Summary of the Invention

[0006] The purpose of this invention is to provide a self-calibrating and adaptive learning intelligent bellows testing system and method to solve the above-mentioned problems existing in the prior art.

[0007] The application is as follows: In a first aspect, the present invention provides a self-calibrating and adaptive learning intelligent bellows testing system, comprising: The basic fluid circulation module is used to provide and circulate a constant-temperature and constant-pressure test fluid to the object under test; The data acquisition module is used to collect the fluid state parameters at the inlet and outlet of the object under test in real time. The reference module is used to provide a known, traceable physical reference for system measurements; The intelligent control module is signal-connected to the basic fluid circulation module, data acquisition module, and reference module; wherein, the intelligent control module is configured to execute: Self-calibration process: The system is controlled to operate under the reference conditions provided by the reference module, and measured data is acquired through the data acquisition module; the theoretical expected value is calculated based on the reference conditions and real-time operating parameters, and the theoretical expected value is compared with the measured data to verify the measurement accuracy of the system; Adaptive learning process: Establish a historical database to store test and calibration data from previous tests, and adaptively optimize the calculation model parameters or generate real-time compensation coefficients for the data acquisition module by analyzing the deviation trends in the historical data.

[0008] Furthermore, the self-calibration process is initiated according to a preset strategy, which is executed by the intelligent control module. This strategy is executed after the system is powered on or when the continuous running time reaches a predetermined period. The intelligent control module is further configured to: If the reference module is determined to be unready when the self-calibration process is scheduled to begin, an interactive reminder will be sent to the user. After the self-calibration process is completed, a self-calibration report is automatically generated and output. This report includes at least the instrument status determination results and measurement deviation data.

[0009] Furthermore, the intelligent control module is configured to have a dynamic environmental compensation function: it introduces real-time collected environmental data to compensate and correct the heat loss model and fluid property parameters in the system in real time, so as to improve the accuracy of theoretical calculations and measured data.

[0010] Furthermore, the intelligent control module is also configured to: continuously monitor the rate of change of the readings of the data acquisition module during self-calibration or testing; when the rate of change is lower than a preset threshold, determine that the system has reached a steady state and trigger a data recording operation.

[0011] Furthermore, the intelligent control module also performs digital filtering on the signals acquired by the data acquisition module to eliminate random noise interference.

[0012] Secondly, the present invention provides a testing method for an intelligent bellows testing system, executed by the intelligent control module, comprising the following steps: Self-calibration steps: Run the self-calibration process and verify the measurement accuracy by comparing the reference theoretical value with the system's measured value; Precise testing procedures: After the accuracy verification is passed, perform performance tests on the bellows under test and record the test data; Adaptive learning steps: Store the test data in the historical database, and optimize the calculation model parameters or generate real-time compensation coefficients based on the historical data trends.

[0013] Furthermore, the method includes an environmental dynamic compensation sub-step in both the self-calibration step and the precision testing step: introducing real-time acquired environmental data to correct the heat loss model and fluid property parameters in theoretical calculations or performance analysis in real time.

[0014] Furthermore, the self-calibration step specifically includes: Dynamic fluid property compensation sub-step: Based on the real-time measured fluid temperature, query or calculate the accurate fluid property parameters at the current temperature; Theoretical value calculation sub-step: Based on the Darcy-Weisbach formula and thermodynamic formula, using the compensated fluid physical property parameters, the known parameters of the standard calibration pipeline and the real-time operating parameters, calculate the theoretical pressure drop value and the theoretical temperature difference value respectively; Deviation analysis and judgment sub-step: Compare the calculated theoretical value with the instrument's measured value. If the deviation exceeds the allowable range, the instrument is judged to be abnormal.

[0015] Furthermore, the adaptive learning step includes: generating a dynamic compensation function by performing regression analysis on the data sequence placed in the historical database, and performing pre-compensation on the real-time collected data.

[0016] Furthermore, the method also includes a report generation step: automatically outputting a test report containing performance parameters and intelligent judgment conclusions.

[0017] Compared with the prior art, the embodiments of the present invention achieve the following beneficial effects: This invention introduces online self-calibration logic. Upon power-on or during a preset cycle, the device automatically verifies and calibrates the accuracy of key sensors (temperature, pressure) using its built-in reference module. This process compares the sensor's measured values ​​with theoretical values ​​calculated based on physical laws and known benchmarks, ensuring the system's measurement accuracy is traceable to the physical reference. It provides online, rapid instrument self-calibration capabilities, guaranteeing test accuracy and ensuring the continuous reliability and consistency of test data.

[0018] The intelligent control unit of this invention dynamically compensates for the effects of environmental changes on fluid properties (density, viscosity) and system heat loss during theoretical value calculations and performance evaluations, making test results under different environmental conditions highly comparable and repeatable. Simultaneously, the built-in steady-state judgment logic automatically captures the moment the system reaches thermal equilibrium, thereby recording the most effective test data and eliminating delays and subjective errors caused by human judgment.

[0019] This invention establishes a historical calibration database and analyzes long-term data trends to automatically identify sensor drift patterns and adaptively optimize calculation model parameters or generate compensation coefficients. It can not only perform real-time software compensation for measured values, but also provide early warnings of the device's own health status, indicating potential instrument degradation or performance decline. This realizes the transformation from passive maintenance to predictive maintenance, greatly improving the long-term reliability and value of the device. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a self-calibration and adaptive learning intelligent bellows testing method according to the present invention. Detailed Implementation

[0021] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments given herein are for illustration and explanation only and are not intended to limit the present invention.

[0022] It should be noted that many specific details are set forth in the following description in order to provide a full understanding of the present invention. However, the present invention may have other embodiments, and therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0023] This embodiment provides a self-calibrating and adaptive learning intelligent bellows testing system, comprising: a basic fluid circulation module, a data acquisition module, a reference module, and an intelligent control module. All modules are centrally controlled and coordinated by an intelligent control module acting as a "nerve center."

[0024] The basic fluid circulation module provides a stable and controllable fluid medium circulation for system testing. Its workflow is as follows: The test medium (usually water) is stored in a water tank and pumped out of the tank. It flows through the heating element and is precisely heated to a set temperature (e.g., 50°C). Subsequently, the thermostatic fluid flows through the inlet test pipe through the test object (corrugated pipe) and finally returns to the water tank via the outlet test pipe and return pipe, forming a closed loop. The core of this module lies in providing precise operating conditions for testing through the coordinated operation of the pump and heating element.

[0025] The data acquisition module is responsible for acquiring key physical quantities during the testing process in real time and with high precision. In this embodiment, a temperature detector and a pressure gauge are integrated on the inlet test pipeline, and a temperature detector and a pressure gauge are integrated on the outlet test pipeline, for synchronously acquiring the temperatures at the inlet and outlet of the bellows. , ) and pressure ( , Data. All sensor signals are transmitted to the intelligent control module.

[0026] In this embodiment, the reference module is implemented using a detachable standard calibration pipe. The pipe's structural parameters (such as length L, inner diameter D, and equivalent inner wall roughness) are specified. All of them have been precisely calibrated and are known. During calibration, they are installed at the test station to provide an absolute reference for theoretical calculations.

[0027] The intelligent control module is the core carrier of the intelligent functions of this invention. It is usually composed of a programmable logic controller (PLC) or an industrial computer, and its internal storage contains control programs that implement various intelligent algorithms. The intelligent control module is connected to all sensors and actuators (pumps, heaters, valves, etc.) via cables, forming a centralized control architecture.

[0028] See attached document Figure 1 As shown, the present invention provides a testing method for an intelligent bellows testing system, comprising the following steps: S1 self-calibration steps: Run the self-calibration process to verify the measurement accuracy by comparing the theoretical reference value with the measured value of the system. S101 Installation of Standard Parts and System Preparation: When the equipment reaches the preset self-calibration time (e.g., after system startup, or when the continuous running time reaches a predetermined cycle), the intelligent control unit will first issue a prompt message to the operator through the human-machine interface (e.g., touch screen), stating "Please install the standard calibration pipeline for system self-calibration." Following the prompt, the operator installs the standard calibration pipeline with known characteristics to the test station. The intelligent control module then starts the basic fluid circulation module, pumping a constant-temperature fluid (e.g., 50°C) into the system and closes the connection valve to the test pipeline, ensuring that calibration is performed within a closed pipeline.

[0029] S102 Thermal Steady-State Determination and High-Precision Data Acquisition: The system enters the thermal steady-state waiting stage; the intelligent control module samples the raw readings of temperature detector one and temperature detector two at a high frequency (e.g., 10Hz), and records them as instantaneous raw values. and .

[0030] right and Digital filtering (preferably using a moving average filtering algorithm) is performed to suppress random noise, resulting in the filtered instantaneous temperature value. and The intelligent control module is based on and Real-time calculation of temperature change rate and When the criterion is met for a continuous period of time (e.g., 30 seconds). and (in When a preset threshold (e.g., 0.1°C / min) is reached, the system is considered to have reached thermal steady state.

[0031] While the intelligent control module determines the thermal steady state, it also collects the readings of pressure gauge 1 over a period of time (e.g., 60 seconds) based on the same filtering process, and calculates and records the average value as follows: Collect the readings of temperature detector one, calculate its average value, and record it as . Collect the readings of pressure gauge 2, calculate its average value, and record it as follows: Collect the readings from temperature detector two, calculate its average value, and record it as follows: Finally, the measured pressure difference was calculated. The measured temperature difference .

[0032] S103 theoretical value calculation and dynamic environmental compensation: Dynamic compensation of fluid properties: The intelligent control module compensates for fluid temperature in real time. The system queries the built-in high-precision water property parameter library based on the IAPWS-IF97 standard and calculates the precise fluid density at the current temperature through interpolation. and dynamic viscosity This sub-step enables dynamic compensation of fluid properties as they change with temperature.

[0033] Theoretical value calculation: The theoretical pressure drop through a standard pipe is calculated using the Darcy-Weisbach formula: in: The coefficient of friction is given by the Körbruck formula. The result was obtained through iterative calculation. For the equivalent roughness of the pipe, It is the Reynolds number; L and D are the known length and inner diameter of the standard calibration pipe; The velocity is estimated from the pump's speed-flow characteristic curve.

[0034] Based on the thermodynamic law of conservation of energy, and incorporating real-time acquired ambient temperature data... Dynamic compensation is applied to the heat loss model. The theoretical temperature difference is calculated using the following formula: in: This represents the power lost due to heat. Here, A represents the overall heat transfer coefficient of the pipeline (determined through prior calibration experiments), and A represents the outer surface area of ​​the pipeline. The average temperature of the fluid (take) and (arithmetic mean) Ambient temperature; For mass flow rate ( (volume flow rate); The specific heat capacity of the fluid is obtained by referring to the property table based on the temperature T.

[0035] S104 Deviation Analysis and Judgment: Calculate the theoretical value... , Real-time measured values ​​of pressure gauge 1, pressure gauge 2, temperature detector 1, and temperature detector 2 , Compare and calculate the relative deviation:

[0036]

[0037] If there is a deviation and If all deviations are less than the preset allowable threshold (e.g., 3%), all relevant instruments are considered normal and calibration is passed. If any deviation exceeds the tolerance, the corresponding instrument is considered abnormal, the system records the deviation data and issues an alarm to guide maintenance.

[0038] S2 Precision Test Procedure: After the system self-calibrates successfully, the performance of the bellows under test is tested.

[0039] S201 Test Initialization and Workpiece Installation. Install the bellows to be tested at the test station and ensure a reliable seal. Start the basic fluid circulation module and introduce the test fluid at the set temperature into the bellows.

[0040] S202 Thermal Steady-State Determination and Formal Data Acquisition. The system monitors the fluid state, and its thermal steady-state determination logic is consistent with that described in the self-calibration process S102. Once the system reaches thermal steady-state, the intelligent control module triggers the acquisition of formal test data. The data acquired at this stage includes the formal inlet temperature. Formal export temperature Formal import pressure and formal export pressure .

[0041] S203: Performance Parameter Calculation and Output. Based on formal test data, the intelligent control module calculates and outputs the bellows' performance parameters. First, the core measured values ​​are calculated: Measured pressure difference Measured temperature difference based on and real-time operating conditions (flow rate) Fluid density Dynamic viscosity ), calculate the flow resistance coefficient that characterizes the flow resistance of the bellows. value:

[0042] based on mass flow rate and specific heat capacity of fluid And introduce ambient temperature As a boundary condition, the heat transfer characteristics (such as heat transfer coefficient) of the bellows are evaluated. value):

[0043] in, Logarithmic mean temperature difference:

[0044] S3 Adaptive Learning Steps: During system operation, it continuously runs in the background, constantly optimizing its judgment benchmarks and output results through in-depth analysis of historical data.

[0045] S301 Historical Data Archiving and Preprocessing: After each self-calibration (S104) or bellows precision test (S203), the intelligent control module automatically packages the complete set of data for this task into a single data unit and archives it with encryption. The data unit includes, but is not limited to: theoretical calculated values ​​(…). , ), Sensor measured values ​​( , ), environmental data ( ), and the calculated performance parameters (K, ), deviation value ( , The system also includes complete timestamps and operational condition tags. Before archiving, the system performs digital filtering (such as moving average filtering) on ​​the raw data for smoothing preprocessing to eliminate the interference of instantaneous noise on long-term trend analysis.

[0046] S302 Multi-dimensional Trend Analysis and Model Parameter Optimization: Multi-dimensional mining of historical databases is performed to identify potential patterns and optimize system models. This is also applied to formal product testing and system self-calibration to achieve continuous optimization.

[0047] The intelligent control unit accesses a historical calibration database, which continuously records theoretical values, measured values, environmental parameters, and deviation data from each self-calibration and test. The intelligent control unit performs multi-dimensional trend analysis, and based on the analysis results, the system automatically generates dynamic compensation coefficients or optimizes model parameters. a) Time dimension analysis: Analyze whether the reading deviation of a specific sensor (such as pressure gauge 1) has a linear or exponential relationship with the operating time, thereby distinguishing between normal aging and abnormal drift.

[0048] The system uses linear regression to analyze historical deviation data. With running time By fitting the data, the baseline drift model is obtained:

[0049] Among them, slope The intercept represents the expected normal aging rate. This represents the initial deviation.

[0050] The system continuously monitors new deviation data points. Calculate its residuals compared to the baseline model:

[0051] Set a threshold based on historical residual statistics (e.g., 2 or 3 standard deviations). If the threshold is exceeded multiple times consecutively (e.g., for three consecutive calibration cycles), it is determined to be "abnormal drift" rather than normal aging.

[0052] Update model parameters and This refines future predictive compensation. It also triggers advanced alerts, indicating "Sensor performance is degrading rapidly; immediate inspection or replacement is recommended," and may mark the sensor as "untrustworthy," reducing its weight in subsequent calculations or discontinuing its use.

[0053] b) Environmental correlation analysis: Analyze whether the temperature sensor deviation is related to the ambient temperature ( There is a strong correlation between the two, thus identifying their temperature drift characteristics and establishing a temperature drift model.

[0054] The system will use sensor deviation data With ambient temperature Regression analysis was performed to fit the temperature drift characteristic curve:

[0055] in This is the temperature drift coefficient. This is the theoretical inherent deviation value of the sensor when the ambient temperature is 0°C.

[0056] Calculate the residual between the actual deviation and the model prediction. .

[0057] The system maintains a residual queue of fixed capacity (e.g., the most recent 20 times) and calculates its mean. ) and standard deviation ( ); When the intelligent control module detects that the residual mean satisfies This indicates that the current temperature drift compensation model has experienced an overall shift. It is the standard deviation of residuals accumulated over a long period during the model's healthy period, and A is the bias coefficient (e.g., 2.0). Although the current model exhibits a bias, its characteristic temperature drift trend (i.e., slope) is still significant. The original model may still be valid. Therefore, the decision is not to immediately discard it, but rather to downgrade it. A temporary, traceable bias compensation is introduced to quickly correct systematic errors and buy time for a fundamental fix.

[0058] The system will use the mean of recent residuals This is calculated as a dynamic bias compensation value. Subsequently, in all measurement calculations relying on this sensor, the raw readings are corrected in real time: in, The raw readings collected by the sensor; This is the correction value after bias compensation.

[0059] The system generates a prompt message on the human-machine interface. This message must accurately indicate the location of the fault and provide clear guidance: "Note: A systematic deviation has occurred in the temperature drift compensation of temperature detector one. Real-time offset compensation has been enabled. The current measurement accuracy can be maintained. The system will automatically collect data in subsequent operations to optimize the model." When the intelligent control module detects that the residual standard deviation meets the requirement This indicates that the sensor's response characteristics or its relationship with ambient temperature have become unstable, leading to a decrease in the reliability of the model's predictions. Here, B is the fluctuation coefficient (e.g., 1.5).

[0060] The system adopts a conservative strategy, reducing reliance on the reliability of current data and triggering in-depth diagnostics. It automatically assigns a wider "measurement uncertainty" range to the measurement results of the sensor (e.g., temperature detector one) and explicitly notes in the final test report that "the result is affected by ambient temperature fluctuations, and its reliability has decreased." A warning is generated: "Warning: The temperature drift characteristics of temperature detector one have become less stable, increasing prediction uncertainty. Measurement results are for reference only; it is recommended to check the sensor status. The system will attempt to automatically relearn the model." Cross-validation analysis: By analyzing the correlation of readings from different sensors under the same operating conditions, system-level deviation diagnosis and collaborative optimization are achieved. It covers two major systems: pressure sensors and temperature sensors, forming a three-dimensional diagnostic network.

[0061] In the pressure sensor correlation analysis system, the system first calculates the inlet pressure deviation and the outlet pressure deviation: Import pressure deviation:

[0062] Export pressure deviation:

[0063] in, and These are the theoretically calculated values ​​corresponding to the inlet and outlet pressures measured during the formal test.

[0064] The system uses a sliding window algorithm to calculate the Pearson correlation coefficient of the reading deviation between pressure sensor 1 and pressure sensor 2 in real time.

[0065] in Describing covariance, It represents the standard deviation.

[0066] The system establishes a multi-level judgment mechanism, when | A correlation coefficient > 0.8 for three consecutive sampling periods is considered strong, indicating a systemic bias; when 0.5 < | If | ≤ 0.8, it is considered a moderate correlation, and enhanced monitoring mode is activated; when | If |≤0.5, it is considered a weak correlation, and routine monitoring should continue.

[0067] In a temperature sensor correlation analysis network, the theoretical temperature difference is calculated in real time. actual temperature difference deviation rate Temperature difference consistency analysis is performed. Simultaneously, the system monitors trend synchronization; when the inlet temperature deviates... Deviation from outlet temperature When the changes are in the same direction and have similar magnitudes, it is determined to be a system-level thermal model deviation; when the deviation shows asynchronous or reverse changes, it is determined to be a sensor unit failure.

[0068] Upon detecting a strong correlation, the system immediately initiates a multi-level collaborative optimization strategy. For system-level parameter optimization, a three-level optimization process is executed: refitting the pump speed-flow characteristic curve parameters to recalibrate the flow model; adjusting the resistance coefficient and pipe resistance parameters of the pressure loss model based on the latest data to optimize the pressure transmission model; and optimizing the weight coefficients of the system-level compensation algorithm according to real-time operating conditions to complete dynamic compensation parameter updates. For adaptive thermal model parameters, the system updates the heat transfer coefficient in real time using the least squares method. Complete online identification of heat transfer coefficient and dynamically adjust specific heat capacity according to temperature changes. and density The calculation parameters are used to optimize fluid properties, and the empirical formulas for optimizing the heat dissipation area A and the convective heat transfer coefficient are used to correct the heat loss model.

[0069] The system establishes a dynamic sensor reliability evaluation system to achieve intelligent weight allocation, and adopts a weight calculation model. in For sensor weights, The standard deviation of recent readings. For the current deviation, , The adjustment coefficient is used. A data fusion strategy is implemented based on the weight calculation results: high-reliability sensor ( >0.8) Dominant measurement results, medium reliability sensor (0.5< ≤0.8) participates in the weighted average, low reliability sensor ( ≤0.5) Temporarily isolate and trigger calibration.

[0070] After each collaborative optimization, the system executes a rigorous closed-loop optimization verification system. Short-term verification includes immediately performing five consecutive measurements to verify the optimization effect, checking the consistency of readings from various sensors, and confirming the stability of system-level parameter updates. Long-term monitoring continuously tracks performance over 24 hours, monitors the changing trends of correlation coefficients, and evaluates the durability of the optimization effect. Through this complete correlation analysis and collaborative optimization mechanism, the system can intelligently identify various sources of deviation and achieve precise parameter optimization. This significantly improves the reliability and stability of the overall testing system.

[0071] The system employs real-time parameter identification based on recursive least squares to ensure timely updates of model parameters; it sets up a rate-of-change constraint mechanism to prevent over-adjustment of parameters; it uses a multi-objective optimization algorithm to balance system accuracy and stability; and it establishes forward verification and rollback safeguards to ensure optimization reliability. After each optimization, the system automatically performs verification tests, confirming the optimization effect by comparing performance metrics before and after optimization.

[0072] The system constructs a complete knowledge management system. After each optimization, the parameter knowledge base is automatically updated, recording the optimization trajectory and effect evaluation. An optimization case library is established, accumulating a complete knowledge chain of "fault phenomenon - optimization solution - effect evaluation". Through machine learning algorithms, the system can learn autonomously from historical optimization cases, continuously improving optimization efficiency. A monthly system health report is generated, automatically identifying optimization blind spots and achieving continuous self-improvement.

[0073] S303 Dual-Loop Optimization: The generated optimization parameters are intelligently allocated to different scenarios, forming two optimization loops: For formal measurements (enhancing the reliability of current results): During performance testing of the bellows under test, the system automatically calls the corresponding dynamic compensation function during the data fusion phase to perform "online cleaning" and "real-time correction" on the raw sensor readings acquired in real time, and then uses them to calculate the final flow resistance coefficient. and heat transfer coefficient This is equivalent to automatically eliminating known systematic errors before the final result is calculated, significantly improving the absolute accuracy and reliability of a single test result.

[0074] Applied to system self-calibration (optimizing future calibration benchmarks): During system self-calibration, the system uses more accurate model parameters (such as...). This allows for the calculation of theoretical values. This optimizes the calibration benchmark itself, enabling the calibration process to not only determine if an instrument is faulty, but also to assess its accuracy on a finer scale. This enhances the sensitivity and foresight of the calibration, thus providing a more reliable standard for all subsequent tests from the outset.

[0075] S304 performance parameter calculation and report generation: Based on highly reliable data that has undergone fusion compensation and adaptive correction, the system performs the final calculation.

[0076] Based on highly reliable data that has been corrected through environmental compensation and historical learning (such as...) , , , (etc.), the system performs the final performance parameter inversion calculation according to the S203 sub-step, and calculates the flow resistance coefficient. and heat transfer coefficient .

[0077] The intelligent control unit has a built-in measurement uncertainty analysis module. Based on error propagation theory, the intelligent control unit automatically analyzes all sources of uncertainty, including the accuracy of each sensor (e.g., pressure gauge accuracy ±0.5%FS, temperature sensor accuracy ±0.1°C), flow estimation error, and model simplification error. It calculates and provides core performance parameters (e.g., ...). , Expanded uncertainty (e.g.) =2.5±0.15 (with a coverage probability of approximately 95%), making the test results more scientific and reliable.

[0078] After the calculation is completed, the intelligent control unit does not simply output the data, but performs intelligent judgment and archiving.

[0079] The calculated performance parameters are compared with product standard thresholds pre-stored in the database. For example, if the calculated flow resistance coefficient... If the value is less than the maximum allowable value specified in the standard, the system will automatically determine that the product is "pass" and mark it with a prominent green mark in the report; otherwise, it will be determined as "fail" and an alarm will be issued with a red mark.

[0080] All raw data, correction processes, final results, environmental data, judgment conclusions, and corresponding product serial numbers, test timestamps, operator IDs, and other information from this test were automatically packaged, encrypted, and stored in a historical test database. This database supports rapid retrieval and statistical analysis by time, product model, test results, etc., providing a data foundation for quality traceability, process improvement, and big data analysis.

[0081] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0082] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. Any of the claimed embodiments can be used in any combination.

Claims

1. A self-calibrating and adaptive learning intelligent bellows testing system, characterized in that, include: The basic fluid circulation module is used to provide and circulate a constant-temperature and constant-pressure test fluid to the object under test; The data acquisition module is used to collect the fluid state parameters at the inlet and outlet of the object under test in real time. The reference module is used to provide a known, traceable physical reference for system measurements; The intelligent control module is signal-connected to the basic fluid circulation module, data acquisition module, and reference module; wherein, the intelligent control module is configured to execute: Self-calibration process: The system is controlled to operate under the reference conditions provided by the reference module, and measured data is acquired through the data acquisition module; the theoretical expected value is calculated based on the reference conditions and real-time operating parameters, and the theoretical expected value is compared with the measured data to verify the measurement accuracy of the system; Adaptive learning process: Establish a historical database to store test and calibration data from previous tests, and adaptively optimize the calculation model parameters or generate real-time compensation coefficients for the data acquisition module by analyzing the deviation trends in the historical data.

2. The intelligent bellows testing system with self-calibration and adaptive learning according to claim 1, characterized in that, The self-calibration process is initiated according to a preset strategy, which is executed by the intelligent control module. This strategy is executed after the system is powered on or when the continuous running time reaches a predetermined period. The intelligent control module is further configured to: If the reference module is determined to be unready when the self-calibration process is scheduled to begin, an interactive reminder will be sent to the user. After the self-calibration process is completed, a self-calibration report is automatically generated and output. This report includes at least the instrument status determination results and measurement deviation data.

3. The intelligent bellows testing system with self-calibration and adaptive learning according to claim 1, characterized in that, The intelligent control module is configured to have a dynamic environmental compensation function: it introduces real-time collected environmental data to compensate and correct the heat loss model and fluid property parameters in the system in real time, so as to improve the accuracy of theoretical calculations and measured data.

4. The intelligent bellows testing system with self-calibration and adaptive learning according to claim 1, characterized in that, The intelligent control module is also configured to: continuously monitor the rate of change of the readings of the data acquisition module during self-calibration or testing; when the rate of change is lower than a preset threshold, determine that the system has reached a steady state and trigger a data recording operation.

5. The intelligent bellows testing system with self-calibration and adaptive learning according to claim 1, characterized in that, The intelligent control module also performs digital filtering on the signals acquired by the data acquisition module to eliminate random noise interference.

6. A testing method based on the intelligent bellows testing system according to any one of claims 1 to 5, characterized in that, Executed by the intelligent control module, the following steps are included: Self-calibration steps: Run the self-calibration process and verify the measurement accuracy by comparing the reference theoretical value with the system's measured value; Precise testing procedures: After the accuracy verification is passed, perform performance tests on the bellows under test and record the test data; Adaptive learning steps: Store the test data in the historical database, and optimize the calculation model parameters or generate real-time compensation coefficients based on the historical data trends.

7. The test method according to claim 6, characterized in that, The method includes, in both the self-calibration step and the precision testing step, an environmental dynamic compensation sub-step: introducing real-time acquired environmental data to correct the heat loss model and fluid property parameters in theoretical calculations or performance analysis in real time.

8. The test method according to claim 6, characterized in that, The self-calibration step specifically includes: Dynamic fluid property compensation sub-step: Based on the real-time measured fluid temperature, query or calculate the accurate fluid property parameters at the current temperature; Theoretical value calculation sub-step: Based on the Darcy-Weisbach formula and thermodynamic formula, using the compensated fluid physical property parameters, the known parameters of the standard calibration pipeline and the real-time operating parameters, calculate the theoretical pressure drop value and the theoretical temperature difference value respectively; Deviation analysis and judgment sub-step: Compare the calculated theoretical value with the instrument's measured value. If the deviation exceeds the allowable range, the instrument is judged to be abnormal.

9. The test method according to claim 6, characterized in that, The adaptive learning steps include: generating a dynamic compensation function by performing regression analysis on the data sequences placed in the historical database, and performing pre-compensation on the real-time collected data.

10. The test method according to claim 6, characterized in that, The method also includes a report generation step: automatically outputting a test report containing performance parameters and intelligent judgment conclusions.

Images