A Smart Test Method for the Comprehensive Performance of High-Voltage Electronic Fans

Abstract

This invention discloses an intelligent testing method for the comprehensive performance of high-voltage electric fans, comprising: establishing a multi-physics coupled test environment of vibration, temperature, and air intake; setting a cyclic test program, with each parameter configured according to the thermal-vibration-fluid coupling failure mechanism; real-time data acquisition; establishing a thermal-vibration-fluid coupling transfer function model, and using model predictive control to adaptively adjust test conditions; extracting multi-dimensional performance degradation features, establishing a degradation trajectory model to extrapolate the remaining service life; and evaluating the comprehensive performance status. This invention significantly improves the consistency between test conditions and actual service, the accuracy of test data, and the life prediction capability through failure mechanism-based condition coupling configuration, adaptive adjustment of model predictive control, degradation trajectory modeling and life prediction, digital twin virtual-real mapping, and robust boundary dynamic search. It achieves an upgrade from qualification judgment to predictive maintenance, providing a comprehensive quantitative basis for the reliability assessment of high-voltage electric fans.

Description

Technical Field

[0001] This invention relates to the technical field of high-voltage electronic fan testing methods, specifically to an intelligent testing method for the comprehensive performance of high-voltage electronic fans. Background Technology

[0002] High-voltage electric fans are core components in new energy vehicle thermal management systems, data center cooling, and industrial ventilation; their reliability directly impacts the safe operation of the entire system. In actual service, high-voltage electric fans not only need to operate continuously in a wide temperature range (e.g., -40℃ to 85℃) but also withstand the intense vibrations generated by vehicle or equipment operation, while ensuring sufficient airflow and air pressure to meet heat dissipation requirements. Therefore, before leaving the factory, high-voltage electric fans must undergo comprehensive performance testing, including but not limited to vibration durability testing, high and low temperature start-up and operation testing, and airflow and noise testing.

[0003] In existing technologies, performance tests for high-voltage electric fans are typically conducted in different independent devices or laboratories. For example, vibration tests need to be performed on a separate vibration table, high and low temperature environment tests need to be performed in an environmental test chamber, and air intake performance tests need to be performed in a wind tunnel or airflow test bench. This decentralized testing mode suffers from drawbacks such as low testing efficiency, insufficient realism in environmental simulation, poor controllability of the testing environment, and insufficient operational safety. To address these issues, some technical solutions have proposed integrated testing devices that physically integrate the vibration table, environmental chamber, and air duct, enabling the application of multi-physics field excitation to the fan under test within the same space, thereby improving testing efficiency to some extent.

[0004] However, even with the aforementioned integrated testing equipment, existing testing methods still suffer from the following deep-seated technical shortcomings: First, the configuration of test conditions lacks support from physical coupling mechanisms. Existing test methods typically set vibration, temperature, and airflow parameters independently based on experience or standards, without fully considering the coupling failure mechanisms between these physical fields. For example, in high-temperature environments, the magnetic properties of the fan magnets demagnetize, and the viscosity of the bearing grease decreases significantly. If high-frequency, high-amplitude vibration excitation is applied under these conditions, a thermal-vibration coupling failure mode will occur that does not match the actual service conditions, resulting in test results that cannot accurately reflect the actual lifespan characteristics of the fan. Conversely, in normal-temperature environments, insufficient vibration excitation cannot effectively induce vibration fatigue failure. Therefore, the lack of a test condition configuration method based on the thermal-vibration-fluid coupling failure mechanism leads to a systematic deviation between test conditions and actual service boundaries.

[0005] Second, the testing process employs an open-loop execution mode, lacking adaptive adjustment capabilities. Existing testing methods mostly execute open-loop cycles according to preset operating condition tables. During testing periods lasting tens or even hundreds of hours, actual test conditions deviate from the target conditions due to factors such as thermal inertia within the test chamber, airflow fluctuations, and changes in vibration table load. Existing methods cannot predictively adjust cooling power, airflow, and vibration excitation based on real-time data acquisition, resulting in poor consistency and repeatability of test conditions, thus affecting the accuracy and reliability of the test data.

[0006] Third, testing and evaluation are limited to performance qualification and lack the ability to predict lifespan. Existing testing methods typically determine whether a fan passes or fails by comparing changes in performance parameters before and after the test. This binary "pass / fail" approach cannot reveal the performance degradation trajectory of a fan during cyclic testing, nor can it extrapolate its remaining lifespan. As high-voltage electric fans evolve towards higher power density and longer lifespan, OEMs' reliability assessment requirements for fans have upgraded from "whether it passes or fails" to "how long it can still be used." Existing testing methods can no longer meet the higher demands of predictive maintenance and reliability assessment.

[0007] Fourth, there is a lack of digital twin mapping between the physical and virtual worlds and the ability to dynamically search for robust boundaries. In existing testing methods, physical testing and virtual simulation are independent of each other, making it impossible to use digital twin models to verify and correct the testing process in real time. At the same time, existing methods mostly use fixed operating condition tables for cyclic testing, which cannot dynamically adjust test parameters according to the real-time performance status of the fan under test to search for its performance failure boundary. This results in wasted testing resources and difficulty in obtaining robustness limit data of the fan.

[0008] In summary, the existing technology lacks a comprehensive intelligent testing method for the performance of high-voltage electronic fans that integrates multi-physics field coupled failure mechanisms, model predictive control adaptive adjustment, degradation trajectory modeling and remaining lifetime prediction, digital twin virtual-real mapping, and robust boundary dynamic search, so as to achieve accurate assessment of the reliability of high-voltage electronic fans throughout their entire life cycle. Summary of the Invention

[0009] To address the shortcomings in the prior art, this invention provides an intelligent testing method for the comprehensive performance of high-voltage electronic fans.

[0010] The technical solution adopted in this invention is: an intelligent testing method for the comprehensive performance of a high-voltage electronic fan, comprising the following steps: S1: Establish a multi-physics field coupling test environment for vibration, temperature and air intake, and fix the high-voltage electronic fan under test on the vibration test table in the heat-insulated and sound-insulated test room; S2: Set up a cyclic test program that includes multiple operating conditions. Each operating condition defines vibration parameters, temperature parameters, air intake parameters and running time. The parameters are configured according to the thermal-vibration-fluid coupling failure mechanism of the actual service conditions of the fan under test, so that the temperature parameters and vibration parameters are inversely coupled, and the air intake parameters and speed parameters are inversely coupled. S3: Start the cycle test and execute each working condition in sequence. The vibration response data, temperature data and air volume data of the fan under test are collected in real time by the sensors set in the test chamber. S4: Establish a thermal-vibration-fluid coupling transfer function model of the fan under test, predict the temperature field and vibration response trend within a preset time period based on the current operating parameters, and adjust the output power of the cooler, the speed of the inlet fan and the driving parameters of the vibration test bench in advance according to the prediction deviation, so that the test conditions dynamically approach the target operating conditions. S5: After completing the preset number of cycles, extract the performance degradation features of the same working conditions in each cycle, construct a multi-dimensional performance degradation feature vector, establish a degradation trajectory model based on the feature vector, and extrapolate the remaining service life of the tested fan. S6: Based on the remaining service life and real-time test data, evaluate the overall performance status of the fan under test.

[0011] Furthermore, the cyclic test program includes five operating conditions, with one cycle lasting 60 minutes, for a total of 72 cycles. The first operating condition is a wind speed of 11 m / s, a cabin temperature of 40°C, a motor speed of 1500 rpm, a running time of 10 minutes, a vibration frequency of 100 Hz, and a vibration acceleration of 24 g. The second operating condition is a wind speed of 16 m / s, a cabin temperature of 40°C, a motor speed of 800 rpm, a running time of 20 minutes, a vibration frequency of 100 Hz, and a vibration acceleration of 24 g. The third operating condition is a wind speed of 11 m / s, a cabin temperature of 40°C, a motor speed of 800 rpm, a running time of 20 minutes, a vibration frequency of 100 Hz, and a vibration acceleration of 24 g. The first operating condition was: wind speed 25 m / s, cabin temperature 40℃, motor speed 800 rpm, operation for 20 minutes, vibration frequency 100 Hz, vibration acceleration 24g; the second operating condition was: wind speed 8 m / s, cabin temperature 85℃, motor speed 2000 rpm, operation for 5 minutes, vibration frequency 20 Hz, vibration acceleration 10g; the third operating condition was: wind speed 6 m / s, cabin temperature 85℃, motor speed 2500 rpm, operation for 5 minutes, vibration frequency 20 Hz, vibration acceleration 10g.

[0012] Furthermore, the configuration logic of the thermal-vibration-fluid coupling failure mechanism is as follows: low vibration frequency and low vibration acceleration are configured for high temperature conditions to avoid magnet demagnetization and bearing lubrication failure caused by thermal-vibration coupling; high vibration frequency and high vibration acceleration are configured for normal temperature conditions to independently assess vibration fatigue failure; low motor speed is configured for high inlet air velocity conditions to simulate low speed and high torque conditions under actual heat dissipation requirements; and high motor speed is configured for low inlet air velocity conditions to simulate high speed and low airflow blockage conditions.

[0013] Furthermore, the thermal-vibration-fluid coupling transfer function model is established through a system identification method. The inputs are cabin temperature, vibration frequency, vibration acceleration, and inlet air velocity, and the outputs are the winding temperature rise of the tested fan, bearing vibration intensity, and airflow attenuation rate. Based on the thermal-vibration-fluid coupling transfer function model, a model predictive control algorithm is used to predict the temperature field and vibration response trend within a preset time period in order to reduce overshoot and suppress transient impacts during operating condition switching.

[0014] Furthermore, during the switching of operating conditions, smooth transition control is performed. Within the preset transition time before the end of the previous operating condition, the gradual change path of each parameter is calculated based on the thermal-vibration-fluid coupling transfer function model, and the vibration parameters, temperature parameters and air intake parameters corresponding to the next operating condition are gradually adjusted so that each parameter reaches the target value at the switching node.

[0015] Furthermore, it also includes anomaly protection steps: during the test, the current, voltage, and temperature rise rate of the fan under test are monitored in real time. When any parameter exceeds the safety threshold, the power supply to the fan under test is automatically cut off and the vibration test bench is stopped. At the same time, the exhaust system is started for emergency heat dissipation. The safety threshold is dynamically adjusted according to the prediction results of the thermal-vibration-fluid coupling transfer function model.

[0016] Furthermore, during the testing process, a digital twin model is established based on the design parameters of the high-voltage electronic fan under test. The vibration response data, temperature data, and air volume data are input into the digital twin model in real time to calculate the deviation between the physical test and the virtual simulation. When the deviation exceeds the preset tolerance range, the target parameters of the corresponding working condition in the cyclic test program are automatically corrected, or an abnormal protection step is triggered.

[0017] Furthermore, the degradation trajectory model is established using support vector regression or long short-term memory network, with the input being the multidimensional performance degradation feature vector and the output being the remaining service life of the tested fan; the multidimensional performance degradation feature vector includes the airflow attenuation rate, vibration harmonic distortion rate, noise spectrum shift, and winding temperature rise drift at the same operating point in each cycle.

[0018] Furthermore, after completing a preset number of basic cycle tests, based on the current performance status of the fan under test and the remaining service life, the vibration parameters and / or temperature parameters of the next cycle are dynamically adjusted through gradient descent or genetic algorithm to gradually approach the performance failure boundary of the fan under test; the combination of multi-physics parameters at the failure boundary is recorded to generate a robustness evaluation spectrum of the fan under test.

[0019] Furthermore, the real-time acquired data is centrally processed through an operation console located outside the thermal insulation and soundproofing test room. The operation console is electrically connected to the vibration test bench, temperature control system, air intake system, and various sensors to achieve remote measurement and control and data acquisition. The operation console integrates a display module, a control module, and a data acquisition module. The data acquisition module is used to simultaneously acquire vibration data, temperature data, and airflow data and perform multi-physical quantity coupling analysis.

[0020] The beneficial effects of this invention are: First, it significantly improves the consistency between test conditions and actual service conditions. This invention incorporates the thermal-vibration-fluid coupling failure mechanism of the tested fan under actual service conditions into the test condition configuration. This couples high-temperature conditions with low vibration parameters, normal-temperature conditions with high vibration parameters, and high intake parameters with low speed parameters. This effectively avoids unrealistic magnet demagnetization and bearing lubrication failure caused by high-frequency, high-vibration conditions under high-temperature environments, as well as missed vibration fatigue failures caused by insufficient vibration excitation under normal-temperature environments. Through targeted condition design based on the failure mechanism, the test conditions can realistically reproduce the multi-physics coupling boundaries of high-voltage electric fans in actual service, significantly improving the authenticity, reliability, and predictive ability of the test data for actual lifespan.

[0021] Secondly, this invention achieves dynamic and precise approximation and overshoot suppression of test conditions. By establishing a thermal-vibration-fluid coupling transfer function model of the fan under test, this invention employs a model predictive control algorithm to predict the temperature field and vibration response trends within a preset time period based on current operating parameters. It then adjusts the cooler output power, inlet fan speed, and vibration test bench drive parameters in advance based on the prediction deviation. Compared to the open-loop execution or post-feedback adjustment of existing technologies, this invention achieves forward-looking and predictive control of multi-physics field test conditions, effectively reducing the overshoot of temperature and vibration parameters, suppressing transient shocks during operating condition switching, ensuring the consistency and repeatability of various operating conditions throughout the 72-cycle test, and significantly improving the accuracy and comparability of test data.

[0022] Third, the testing and evaluation are upgraded from "pass / fail judgment" to "lifespan prediction." After completing a preset number of cycles, this invention extracts multi-dimensional performance degradation features such as airflow attenuation rate, vibration harmonic distortion rate, noise spectrum shift, and winding temperature rise drift at the same operating point in each cycle. A performance degradation feature vector is constructed, and a degradation trajectory model is established based on support vector regression or long short-term memory networks to extrapolate the remaining lifespan of the tested fan. Compared to existing technologies that can only provide a binary "pass / fail" result, this invention can reveal the performance degradation trajectory of the fan during cyclic testing, providing a quantitative basis for reliability assessment, maintenance strategy formulation, and warranty cycle design of high-voltage electronic fans, meeting the higher demands of OEMs for predictive maintenance and full life-cycle reliability management.

[0023] Fourth, real-time verification and correction of the testing process are achieved through digital twin mapping. This invention establishes a digital twin model based on the design parameters of the high-voltage electric fan under test, inputs physical test data into the digital twin model in real time, calculates the deviation between the physical test and the virtual simulation, and automatically corrects the target parameters for the corresponding operating condition or triggers anomaly protection when the deviation exceeds a preset tolerance range. This virtual-real mapping mechanism effectively identifies abnormal deviations and potential faults during the testing process, prevents the accumulation of invalid test data, and improves the utilization efficiency of test resources and the controllability of the testing process.

[0024] Fifth, robust boundary dynamic search yielded the fan's performance limit data. After completing the basic cyclic test, based on the current performance state and remaining service life of the tested fan, this invention dynamically adjusts the vibration parameters and / or temperature parameters for the next cycle using gradient descent or a genetic algorithm, gradually approaching the performance failure boundary of the tested fan, and recording the multi-physics parameter combinations at the failure boundary to generate a robustness evaluation spectrum. Compared with existing technologies that use fixed operating condition tables for cyclic testing, this invention can intelligently adjust the testing strategy according to the individual differences and real-time status of the tested fan, avoiding the waste of testing resources, and obtaining robustness limit data that cannot be obtained by existing methods, providing more comprehensive performance boundary information for fan design optimization and selection.

[0025] Sixth, it achieves a smooth transition between operating conditions and improves test safety. During operating condition switching, this invention calculates the gradual change path of each parameter based on a thermal-vibration-fluid coupling transfer function model. Within a preset transition time, it gradually adjusts the vibration, temperature, and air intake parameters corresponding to the next operating condition, avoiding mechanical and thermal shocks caused by sudden parameter changes to the fan under test. Simultaneously, by real-time monitoring of the fan's current, voltage, and temperature rise rate, and dynamically adjusting the safety threshold based on the prediction results of the thermal-vibration-fluid coupling transfer function model, it automatically cuts off the power and initiates emergency cooling when parameters exceed the safety threshold, effectively protecting the tested sample and testing equipment and improving the safety and reliability of the testing process.

[0026] In summary, this invention integrates multi-physics coupled failure mechanisms, adaptive adjustment of model predictive control, degradation trajectory modeling and remaining lifetime prediction, digital twin virtual-real mapping, and robust boundary dynamic search, upgrading the traditional decentralized, open-loop, and decision-based testing methods for high-voltage electronic fans into an integrated, closed-loop, and predictive intelligent testing method. This significantly improves testing efficiency, data accuracy, lifetime prediction capabilities, and testing resource utilization, providing strong technical support for the reliability verification and quality control of high-voltage electronic fans.

[0027] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the accompanying drawings. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the structure of the present invention.

[0029] Figure 2 This is a schematic diagram from another perspective of the present invention.

[0030] Figure 3 This is a schematic diagram of the internal structure of the present invention.

[0031] Figure 4 This is a cross-sectional schematic diagram of the present invention.

[0032] Figure 1-4 The components include: 1. Thermal insulation and soundproofing test chamber; 2. Vibration test bench; 3. Tooling; 4. High-pressure electronic fan; 5. Condenser; 6. Cooler; 7. Air inlet duct; 8. Air inlet fan; 9. Louvered air valve; 10. Heating unit; 11. Control console; 12. Thermal insulation and soundproofing door; 13. Exhaust vent; 14. Silencing louvered air valve. Detailed Implementation

[0033] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0034] Example 1 Reference Figures 1 to 4 The device structure shown in this embodiment provides a smart testing method for the comprehensive performance of a high-voltage electronic fan, specifically including the following steps: S1: Establish a multi-physics coupled test environment for vibration, temperature, and air intake. The high-voltage electronic fan 4 under test is fixed to the vibration test bench 2 inside the thermal insulation and soundproof test chamber 1 using a special tooling 3. In this embodiment, the thermal insulation and soundproof test chamber 1 includes an interior panel, a thermal insulation layer, and a sound insulation layer arranged sequentially from the inside to the outside. Thermal insulation and soundproof doors 12 are respectively provided at both ends of the test chamber 1. The thermal insulation and soundproof door 12 includes a door frame, a door leaf, and a sealing strip provided on the edge of the door leaf. The inside of the door leaf is filled with thermal insulation and soundproof material. The vibration test bench 2 is a high-frequency vibration test bench capable of providing an acceleration of not less than 25G, and is located in the central area inside the test chamber 1. The temperature control system includes multiple condensers 5 installed inside the test chamber 1, and a cooling machine 6 installed outside the thermal insulation and soundproof test chamber 1 and connected to each of the condensers 5 via pipelines. The air intake system includes an air intake duct 7 located on one side of the high-voltage electronic fan 4 under test. The air inlet of the air intake duct 7 is at a preset distance from the air intake end of the high-voltage electronic fan 4 under test. This preset distance is preferably 0.5-2 meters, and 1.2 meters is selected in this embodiment. An air intake fan 8 and a louvered air valve 9 are arranged on the air intake duct 7 along the air intake direction. A heating unit 10 and a cooling unit 15 are connected to the air inlet of the air intake duct 7. The exhaust system includes an exhaust vent 13 located at the top of the test chamber 1. The exhaust vent 13 is equipped with a sound-absorbing louvered air valve 14. Temperature sensors, wind speed sensors, noise sensors, and vibration acceleration sensors are also arranged inside the test chamber 1. Each sensor is electrically connected to an operation control console 11 located outside the thermal insulation and soundproof test chamber 1. The operation control console 11 includes a display module, a control module, and a data acquisition module, and is electrically connected to the vibration test bench 2, the temperature regulation system, the air intake system, and the exhaust system.

[0035] S2: Set up the cyclic test program and configure the thermal-vibration-fluid coupling failure mechanism. Configure a cyclic test program with five operating conditions in the operation console 11. Each cycle lasts 60 minutes, and the program will run a total of 72 cycles. The parameters for each operating condition are configured as follows: Condition 1: Wind speed 11 m / s, cabin temperature 40℃, motor speed 1500 rpm, operation for 10 minutes, vibration frequency 100 Hz, vibration acceleration 24g; Condition 2: Wind speed 16 m / s, cabin temperature 40℃, motor speed 800 rpm, operation for 20 minutes, vibration frequency 100 Hz, vibration acceleration 24g; Condition 3: Wind speed 25 m / s, cabin temperature 40℃, motor speed 800 rpm, operation for 20 minutes, vibration frequency 100 Hz, vibration acceleration 24g; Condition 4: Wind speed 8 m / s, cabin temperature 85℃, motor speed 2000 rpm, operation for 5 minutes, vibration frequency 20 Hz, vibration acceleration 10g; Condition 5: Wind speed 6 m / s, cabin temperature 85℃, motor speed 2500 rpm, operation for 5 minutes, vibration frequency 20 Hz, vibration acceleration 10g.

[0036] The parameters mentioned above are configured according to the thermal-vibration-fluid flow coupling failure mechanism of the tested fan under actual service conditions. Specifically, the configuration logic is as follows: Under normal temperature conditions (conditions 1-3), a high vibration frequency of 100Hz and a high vibration acceleration of 24g are configured to independently assess the fan's vibration fatigue failure mode under normal temperature conditions, fully stimulating potential defects such as structural stress, weld fatigue, and loose electrical connections. Under high temperature conditions (conditions 4 and 5), a low vibration frequency of 20Hz and a low vibration acceleration of 10g are configured to avoid accelerated failure due to thermal-vibration coupling caused by demagnetization of the magnet and decrease in bearing grease viscosity under high temperature conditions, ensuring that the high-temperature test primarily assesses the independent contribution of thermal aging failure. Simultaneously, regarding the coupling of inlet air parameters and rotational speed, a low motor speed of 800rpm is configured for the high airflow speed (25m / s) condition to simulate the low-speed, high-torque condition under actual heat dissipation requirements; a high motor speed of 2500rpm is configured for the low airflow speed (6m / s) condition to simulate the high-speed, low-airflow congestion condition. Through the above coupling configuration, the test conditions can be realistically reproduced to reflect the multiphysics boundary conditions of the tested fan in actual service.

[0037] S3: Start the loop test and collect data in real time. Initiate the cyclic test and execute each operating condition sequentially according to the cyclic test procedure. During the test, the ambient temperature of the chamber and the surface temperature of the fan windings are collected in real time by a temperature sensor installed in test chamber 1, the intake air speed is collected by a wind speed sensor, the sound pressure level data is collected by a noise sensor, the vibration response data of the fan housing and bearing housing is collected by a vibration acceleration sensor, and the input current and voltage of the fan are collected by current and voltage sensors. All sensor data are transmitted to the data acquisition module of the operation console 11 at a sampling frequency of not less than 100Hz for synchronous acquisition and timestamp alignment.

[0038] S4: Establish a thermal-vibration-fluid coupling transfer function model and execute model predictive control adaptive adjustment. During the initial calibration phase, an excitation signal covering the operating conditions is applied to the fan under test, and multiple sets of input-output data are collected. The inputs include cabin temperature T_cabin, vibration frequency f_vib, vibration acceleration a_vib, and inlet air velocity v_wind; the outputs include fan winding temperature rise ΔT_winding, bearing vibration intensity V_rms, and airflow attenuation rate Q_decay. Based on the above data, a thermal-vibration-fluid coupling transfer function model of the fan under test is established using a subspace identification method (such as the N4SID algorithm). The model is in discrete state-space form. x(k+1) = A·x(k) + B·u(k); y(k) = C·x(k) + D·u(k); Where u(k) = [T_cabin(k), f_vib(k), a_vib(k), v_wind(k)]^T is the input vector at time k, y(k) = [ΔT_winding(k), V_rms(k), Q_decay(k)]^T is the output vector, x(k) is the state vector, and A, B, C, and D are the system matrices obtained through system identification.

[0039] During the online execution phase of the cyclic testing, the Model Predictive Control (MPC) algorithm is adopted based on the above transfer function model. The prediction time domain is set to Np = 20 steps and the control time domain to Nc = 5 steps. At each sampling time k, based on the current state x(k) and the model, the output sequence ŷ(k+i|k) within the next Np steps is predicted. The objective function is constructed as follows: J = Σ_{i=1}^{Np} ||ŷ(k+i|k) - y_ref(k+i)||^2_Q + Σ_{i=0}^{Nc-1} ||Δu(k+i)||^2_R Where y_ref is the target operating condition parameter vector, Q and R are weight matrices, and Δu is the control increment. The optimal control sequence is obtained by solving a constrained quadratic programming optimization problem, and the first step of the control is implemented. The control quantities include the output power P_cool of the cooler 6, the rotational speed n_fan of the intake fan 8, and the driving voltage U_vib of the vibration test bench 2, with constraints P_cool∈[0, P_max], n_fan∈[0, n_max], and U_vib∈[0, U_max]. Through this model predictive control, the parameters of each system are adjusted in advance, so that the test conditions dynamically approximate the target operating condition and effectively reduce overshoot.

[0040] S5: Smooth transition control during operating condition switching During the switching of operating conditions, smooth transition control is implemented. Starting from the preset transition time before the end of the previous operating condition (preferably 30 seconds in this embodiment), the optimal gradual change path from the current operating condition parameters to the target parameters of the next operating condition is calculated based on the thermal-vibration-fluid coupling transfer function model. Specifically, the S-curve planning method is adopted to ensure that each parameter changes gradually according to a continuously differentiable velocity-acceleration curve within the transition time, avoiding mechanical and thermal shocks caused by abrupt parameter changes to the tested fan, and ensuring that each parameter accurately reaches the target value at the switching node.

[0041] S6: Digital Twin Virtual-Real Mapping A digital twin model is established based on the design parameters of the tested high-voltage electric fan 4 (including but not limited to the grade and remanence parameters of the magnet, the number of turns and resistance of the stator winding, the bearing type and clearance, and the blade geometry and aerodynamic parameters). The digital twin model is an integrated multiphysics simulation model, including an electromagnetic field finite element model, a thermal network model, and a fluid-structure interaction vibration model. During the test, vibration response data, temperature data, and airflow data collected from the physical test are input into the digital twin model in real time to calculate the virtual simulation output value y_virtual under the current operating condition. The deviation δ between the physical test value y_physical and the virtual simulation value y_virtual is calculated as δ = |y_physical - y_virtual| / |y_virtual|. When the deviation δ exceeds a preset tolerance range (preferably 5% in this embodiment), an abnormal deviation is determined between the physical test and the virtual simulation. The control console 11 automatically corrects the target parameters of the corresponding operating condition in the cyclic test program (e.g., lowering the target temperature by 5% to correct thermal inertia deviation) or triggers an abnormal protection step.

[0042] S7: Anomaly Protection During the test, the operation console 11 monitors the current I, voltage U, and winding temperature rise rate dT_winding / dt of the fan under test in real time. The safety threshold is not a fixed constant but is dynamically adjusted based on the prediction results of the thermal-vibration-fluid coupling transfer function model. Specifically, under the current operating conditions, the upper limit of the normal temperature rise rate predicted by the model is (dT / dt)_pred, so the dynamic safety threshold is set to (dT / dt)_limit = 1.2 × (dT / dt)_pred. When any monitored parameter exceeds the corresponding dynamic safety threshold, the operation console 11 automatically cuts off the power supply to the fan under test and stops the vibration test bench 2. Simultaneously, it opens the silencer louver damper 14 of the exhaust system for emergency heat dissipation, protecting the tested sample and test equipment.

[0043] S8: Degradation Trajectory Modeling and Remaining Useful Life Extrapolation After each cycle is completed, the performance data of each operating point within that cycle is extracted from the database of the operation console 11, and the multidimensional performance degradation characteristics are calculated: Airflow attenuation rate: Q_decay(i) = (Q_1 - Q_i) / Q_1 × 100% Vibration harmonic distortion rate: THD_vib(i) = √(Σ_{h=2}^{H} A_h^2) / A_1 × 100% Noise spectrum offset: Δf_noise(i) = f_peak(i) - f_peak(1) Winding temperature rise drift: ΔT_wind(i) = T_winding(i) - T_winding(1) Where i is the cycle number, Q_1, A_1, f_peak(1), and T_winding(1) are the reference values ​​under the working conditions corresponding to the first cycle, and A_h is the amplitude of the h-th harmonic of the vibration signal.

[0044] Construct a multidimensional performance degradation feature vector X_i = [Q_decay(i), THD_vib(i), Δf_noise(i), ΔT_wind(i)]^T.

[0045] A degradation trajectory model is established based on the feature vector sequence {X_1, X_2, ..., X_i}. In this embodiment, a Long Short-Term Memory (LSTM) network is preferably used to establish the degradation trajectory model. The LSTM network consists of two LSTM layers, each containing 64 hidden units, and a fully connected output layer. The input is the feature vector sequence of the first i cycles, and the output is the remaining service life (RUL) of the tested fan. The network is pre-trained using a degradation dataset of historical fans of the same model, and fine-tuned online using real-time data of the current tested fan during testing. Alternatively, when the amount of data is limited, the Support Vector Regression (SVR) method can also be used, with the RBF kernel selected as the kernel function, and the penalty parameter C and kernel parameter γ determined through grid search and cross-validation.

[0046] When the extrapolated remaining service life RUL is lower than the preset safe service life threshold (preferably 20% of the initial predicted service life in this embodiment), it is determined that the tested fan has a risk of performance degradation.

[0047] S9: Robust Boundary Dynamic Search After completing 72 basic cycle tests, if the remaining service life (RUL) of the tested fan is still higher than the safe service life threshold, the robust boundary dynamic search phase begins. Based on the current performance state and remaining service life of the tested fan, the vibration parameters and / or temperature parameters for the next cycle are dynamically adjusted through an optimization algorithm to gradually approach the performance failure boundary of the tested fan.

[0048] In this embodiment, a genetic algorithm is preferably used for dynamic search. Chromosomes are encoded as a four-parameter vector [T_cabin, f_vib, a_vib, v_wind]. The fitness function is defined as the inverse of the performance degradation rate (i.e., maximizing the degradation rate to accelerate failure), while constraining each parameter to not exceed a safety threshold. Iterative evolution is performed through selection, crossover, and mutation operations until the fan performance parameters exceed the acceptable range. The corresponding multiphysics parameter combination at this point is recorded as the failure boundary point. Alternatively, a gradient descent method can be used, gradually increasing the excitation intensity along the gradient direction of the vibration acceleration with respect to the performance degradation feature vector until the failure boundary is reached.

[0049] By recording multiple failure boundary points and fitting them in a four-dimensional parameter space, a robustness evaluation map of the tested fan is generated, providing data support for fan design optimization and extreme condition applications.

[0050] S10: Comprehensive Performance Status Assessment The operation console 11 summarizes the full-cycle data from 72 cyclic tests, the predicted remaining useful life (RUL), the robustness assessment graph, and the digital twin virtual-real mapping deviation records, generating a comprehensive performance evaluation report for the tested high-voltage electric fan. The report includes: performance stability evaluation under various operating conditions, degradation trend curves, remaining useful life prediction ranges, robustness boundary parameters, and a degree of agreement with the digital twin model. This provides comprehensive and accurate quantitative basis for determining the reliability level of the tested fan, establishing warranty periods, and optimizing maintenance strategies.

[0051] Example 2 The difference between this embodiment and Embodiment 1 is that in step S8, the degradation trajectory model uses Support Vector Regression (SVR) instead of a Long Short-Term Memory network. Specifically, the multidimensional performance degradation feature vector X_i of each cycle is used as the input feature of SVR, and the number of completed cycles is used as the surrogate label. The degradation trajectory curve is fitted by Support Vector Regression and extrapolated to the performance failure threshold to calculate the remaining lifetime. The remaining steps are the same as in Embodiment 1 and will not be repeated.

[0052] Example 3 The difference between this embodiment and Embodiment 1 is that in step S9, the robust boundary dynamic search uses gradient descent instead of the genetic algorithm. Specifically, with vibration acceleration a_vib as the optimization variable and the magnitude ||X|| of the performance degradation feature vector as the optimization objective, the vibration acceleration is gradually increased along the direction of gradient ∇||X|| / ∇a_vib, with each increment being 0.5g, until the fan airflow attenuation rate exceeds 10% or the winding temperature rise exceeds the insulation class limit. The vibration acceleration at this point is recorded as the robust boundary. The remaining steps are the same as in Embodiment 1 and will not be repeated.

[0053] Please note to all technical personnel: Although the present invention has been described according to the specific embodiments above, the ideas of the present invention are not limited to this invention. Any modifications that utilize the ideas of the present invention will be included within the scope of protection of this patent.

Claims

1. A smart testing method for the comprehensive performance of a high-voltage electronic fan, characterized in that, Includes the following steps: S1: Establish a multi-physics field coupling test environment for vibration, temperature and air intake, and fix the high-voltage electronic fan under test on the vibration test table in the heat-insulated and sound-insulated test room; S2: Set up a cyclic test program that includes multiple operating conditions. Each operating condition defines vibration parameters, temperature parameters, air intake parameters and running time. The parameters are configured according to the thermal-vibration-fluid coupling failure mechanism of the actual service conditions of the fan under test, so that the temperature parameters and vibration parameters are inversely coupled, and the air intake parameters and speed parameters are inversely coupled. S3: Start the cycle test and execute each working condition in sequence. The vibration response data, temperature data and air volume data of the fan under test are collected in real time by the sensors set in the test chamber. S4: Establish a thermal-vibration-fluid coupling transfer function model of the fan under test, predict the temperature field and vibration response trend within a preset time period based on the current operating parameters, and adjust the output power of the cooler, the speed of the inlet fan and the driving parameters of the vibration test bench in advance according to the prediction deviation, so that the test conditions dynamically approach the target operating conditions. S5: After completing the preset number of cycles, extract the performance degradation features of the same working conditions in each cycle, construct a multi-dimensional performance degradation feature vector, establish a degradation trajectory model based on the feature vector, and extrapolate the remaining service life of the tested fan. S6: Based on the remaining service life and real-time test data, evaluate the overall performance status of the fan under test.

2. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: The cyclic test program includes five operating conditions, with one cycle lasting 60 minutes, and a total of 72 cycles are executed. The first operating condition is: wind speed 11 m / s, chamber temperature 40℃, motor speed 1500 rpm, running for 10 minutes, vibration frequency 100 Hz, and vibration acceleration 24g. The second operating condition is: wind speed 16 m / s, chamber temperature 40℃, motor speed 800 rpm, running for 20 minutes, vibration frequency 100 Hz, and vibration acceleration 24g. The third operating condition is: wind speed 25 m / s, chamber temperature 40℃, motor speed 800 rpm, running for 20 minutes, vibration frequency 100 Hz, and vibration acceleration 24g. The fourth operating condition is: wind speed 8 m / s, chamber temperature 85℃, motor speed 2000 rpm, running for 5 minutes, vibration frequency 20 Hz, and vibration acceleration 10g. The fifth operating condition is: wind speed 6 m / s, chamber temperature 85℃, motor speed 2500 rpm, running for 5 minutes, vibration frequency 20 Hz, and vibration acceleration 10g.

3. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: The configuration logic of the thermal-vibration-fluid coupling failure mechanism is as follows: low vibration frequency and low vibration acceleration are configured for high temperature conditions to avoid magnet demagnetization and bearing lubrication failure caused by thermal-vibration coupling; high vibration frequency and high vibration acceleration are configured for normal temperature conditions to independently assess vibration fatigue failure; and low motor speed is configured for high air intake speed conditions to simulate low speed and high torque conditions under actual heat dissipation requirements. The low air intake velocity condition is configured with a high motor speed to simulate a high-speed, low-airflow, blockage condition.

4. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: The thermal-vibration-fluid coupling transfer function model is established through a system identification method. The inputs are cabin temperature, vibration frequency, vibration acceleration, and inlet air velocity, and the outputs are the winding temperature rise of the tested fan, bearing vibration intensity, and airflow attenuation rate. Based on the thermal-vibration-fluid coupling transfer function model, a model predictive control algorithm is used to predict the temperature field and vibration response trend within a preset time period in order to reduce overshoot and suppress transient impacts during operating condition switching.

5. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: When switching operating conditions, smooth transition control is performed. Within the preset transition time before the end of the previous operating condition, the gradual change path of each parameter is calculated based on the thermal-vibration-fluid coupling transfer function model. The vibration parameters, temperature parameters and air intake parameters corresponding to the next operating condition are gradually adjusted so that each parameter reaches the target value at the switching node.

6. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: It also includes anomaly protection steps: during the test, the current, voltage and temperature rise rate of the fan under test are monitored in real time. When any parameter exceeds the safety threshold, the power supply of the fan under test is automatically cut off and the vibration test bench is stopped. At the same time, the exhaust system is started for emergency heat dissipation. The safety threshold is dynamically adjusted according to the prediction results of the thermal-vibration-fluid coupling transfer function model.

7. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: During the test, a digital twin model is established based on the design parameters of the high-voltage electronic fan under test. The vibration response data, temperature data, and air volume data are input into the digital twin model in real time to calculate the deviation between the physical test and the virtual simulation. When the deviation exceeds the preset tolerance range, the target parameters of the corresponding working condition in the cyclic test program are automatically corrected, or an abnormal protection step is triggered.

8. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: The degradation trajectory model is established using support vector regression or long short-term memory network. The input is the multidimensional performance degradation feature vector, and the output is the remaining service life of the fan under test. The multidimensional performance degradation feature vector includes the air volume attenuation rate, vibration harmonic distortion rate, noise spectrum shift, and winding temperature rise drift at the same operating point in each cycle.

9. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: After completing a preset number of basic cycle tests, based on the current performance status of the fan under test and its remaining service life, the vibration parameters and / or temperature parameters of the next cycle are dynamically adjusted through gradient descent or genetic algorithm to gradually approach the performance failure boundary of the fan under test; the combination of multi-physics parameters at the failure boundary is recorded to generate a robustness evaluation spectrum of the fan under test.

10. The intelligent testing method for the comprehensive performance of a high-voltage electronic fan according to claim 1, characterized in that: The real-time acquired data is centrally processed through an operation console located outside the thermal insulation and soundproofing test room. The operation console is electrically connected to the vibration test bench, temperature control system, air intake system, and various sensors to achieve remote measurement and control and data acquisition. The operation console integrates a display module, a control module, and a data acquisition module. The data acquisition module is used to simultaneously acquire vibration data, temperature data, and air volume data and perform multi-physical quantity coupling analysis.

Images