Test Method for Failure Causes of Heat Exchangers Based on Digital Filtering and Virtual Damage

Abstract

The present invention relates to the technical field of reliability test and analysis of automotive heat exchangers, and discloses a test method for the failure cause of a heat exchanger based on digital filtering and virtual damage, which includes the following steps: sensor arrangement; starting the engine for testing to obtain the acceleration time-domain signal and the strain signals of each measuring point; through filtering and pseudo-damage calculation of the strain signal ε of each measuring point, obtaining the thermal fatigue cycle times and vibration fatigue cycle times corresponding to the standard amplitude of the test strain load of the same virtual cycle, selecting the maximum value for numerical comparison, and finding out the main cause of the heat exchanger failure during this engine test. The test method for the failure cause of the heat exchanger based on digital filtering and virtual damage in the present invention can quickly, efficiently and accurately analyze and evaluate the virtual damage of various fatigue loads of the heat exchanger on the engine test bench and conduct quantitative comparison to clarify the root cause of the failure.

Description

Technical Field

[0001] The present invention relates to the technical field of reliability test and analysis of automotive heat exchangers, and particularly relates to a test method for the failure cause of a heat exchanger based on digital filtering and virtual damage. Background Art

[0002] Automotive heat exchangers undertake the heat exchange function of the automotive thermal management system. According to the connection situation with the engine, they can be divided into two categories: The first category is heat exchangers directly and rigidly connected to the engine, such as exhaust gas recirculation coolers (EGR), indirect intercoolers, oil coolers, etc.; The second category is heat exchangers not directly and rigidly connected to the engine, such as radiators, air-to-air intercoolers, condensers, evaporators, etc.

[0003] For the first category of heat exchangers, since they are rigidly connected to the engine and directly affected by the vibration excitation load during the engine operation, various assessment tests that must be experienced during engine development need to be installed on the engine. During the engine test, due to the combined complex loads brought by vibration, temperature, and pressure, it is often difficult to quickly and accurately find the root cause and type of failure when the heat exchanger fails during the test, which brings difficulties to product improvement for engineering technicians.

[0004] In order to find the root cause of the failure, it is often necessary to design separate test schemes for the vibration acceleration load, temperature load, and pressure load generated during the engine test. Then, separate test verifications are carried out on the vibration test bench, temperature alternating test bench, and pressure alternating test bench for the acceleration power spectral density (PSD), temperature load change, and pressure load change during the test respectively. Except for the pressure load, the above verification processes have a long cycle, high cost, complex verification processes, non-intuitive verification methods, and poor correspondence. Summary of the Invention

[0005] The purpose of the present invention is to address the deficiencies of the above technologies, and provide a test method for the failure cause of a heat exchanger based on digital filtering and virtual damage, which can quickly, efficiently, and accurately analyze and evaluate the virtual damage of various fatigue loads of the heat exchanger on the engine test bench and perform quantitative comparison to clarify the root cause of the failure.

[0006] To achieve the above purpose, the test method for the failure cause of a heat exchanger based on digital filtering and virtual damage designed by the present invention includes the following steps:

[0007] A) Sensor arrangement: Through finite element analysis of the heat exchanger and combining with the after-sales failure experience of similar products, evaluate the working conditions of the heat exchanger. Select the stress concentration areas and failure-prone areas of the heat exchanger as the measuring points to arrange strain gauges, measure the strain signal ε of each measuring point during engine testing, and arrange three-direction vibration acceleration sensors on the heat exchanger shell to measure the acceleration time-domain signal g of the heat exchanger during testing;

[0008] B) Start the engine for testing to obtain the acceleration time-domain signal g and the strain signal ε of each measuring point;

[0009] C) By filtering the strain signal ε of each measuring point of the heat exchanger during a stable engine test cycle and performing pseudo-damage calculation, obtain the standard amplitude ε pseudo of the thermal fatigue cycle number N T and the vibration fatigue cycle number N V , select the maximum value N T of the thermal fatigue cycle number N Tmax and the maximum value N V of the vibration fatigue cycle number N Vmax for numerical comparison. If N Vmax > N Tmax , then vibration fatigue is the main cause of heat exchanger failure during this engine test. Otherwise, thermal fatigue is the main failure cause. The measuring points corresponding to N Vmax and N Tmax are the vibration fatigue weak point and the thermal fatigue weak point respectively.

[0010] Preferably, in the step C), the thermal fatigue load pseudo-damage calculation includes the following steps:

[0011] C1) Perform Butterworth low-pass filtering on the strain signal ε of each measuring point of the heat exchanger during a stable engine test cycle. The filtering cut-off frequency is 10 hz, and the filtering type is selected to retain the phase. The obtained strain time-domain signal ε LP is used as the thermal fatigue load time-domain signal ε LP of the heat exchanger during the engine test cycle. The pulsating amplitude of ε increases as the engine speed increases or approaches the critical speed, and the signal generally has no symmetry. The pulsating amplitude of ε LP no longer increases as the engine speed increases or approaches the critical speed;

[0012] C2) Set the resolution to perform rainflow counting method analysis on the thermal fatigue load time-domain signal ε LP obtained in the step C1), calculate the occurrence times of different amplitude strain changes based on the resolution, and generate the thermal fatigue load time-domain signal ε LPFor the rain flow matrix, assume the slope of the S-N fatigue life curve of the product is k. When the heat exchanger is in the engine test cycle, the amplitude of the thermal fatigue load change is ε iLP The occurrence times is n iLP , i is the resolution, ε pseudo is the standard amplitude of the virtual cyclic test strain load. Then the pseudo-damage of the thermal fatigue load is equivalent to the standard amplitude ε of the virtual cyclic test strain load pseudo After times of damage caused, use Miner Rule for the time-domain signal ε of the thermal fatigue load of the heat exchanger LP Accumulate all the stress change amplitudes and their occurrence times in the rain flow matrix of ε pseudo The number of test times corresponding to ε pseudo to obtain the thermal fatigue cycle times with the same fatigue pseudo-damage amplitude of ε j is the upper limit of the resolution.

[0013] Preferably, in the step C), the calculation of the pseudo-damage of the vibration fatigue load includes the following steps:

[0014] C3) Perform Butterworth high-pass filtering on the strain signal ε of each measuring point, with the filtering cut-off frequency of 10 hz, and the filtering type selects to retain the phase. The obtained strain time-domain signal ε after filtering HP , ε HP The pulsation amplitude of increases with the increase of the engine speed or approaching the critical speed, and has strong symmetry;

[0015] C4) Perform fast Fourier transform on the strain time-domain signal ε HP to obtain the vibration strain frequency spectrum diagram, and perform fast Fourier transform on the acceleration time-domain signal g to obtain the acceleration frequency spectrum diagram. Check the change trends and extreme frequencies of the vibration strain frequency spectrum diagram and the acceleration frequency spectrum diagram. If the change trends and extreme frequencies of the two are the same, then use the strain time-domain signal ε HP as the vibration fatigue load time-domain signal ε of the heat exchanger during the engine test cycle HP , if they are inconsistent, adjust the filtering cut-off frequency of the high-pass filtering, correct the pressure load of the strain signal ε of the heat exchanger, and analyze and eliminate the noise of the strain signal ε until the change trends and extreme frequencies of the two are the same;

[0016] C5) Set the resolution and perform rain flow counting method analysis on the vibration fatigue load time-domain signal ε obtained in the step C4 HP respectively, calculate and count the occurrence times of different amplitude strain changes based on the resolution, and generate the rain flow matrix of the vibration fatigue load time-domain signal ε HP Assume the slope of the S-N fatigue life curve of the product is k. When the heat exchanger is in the engine test cycle, the amplitude of the vibration fatigue load change is ε iHPThe occurrence times is n iHP , i is the resolution, ε pseudo is the standard amplitude of the virtual cyclic test strain load, then the vibration fatigue load damage is equivalent to the standard amplitude ε of the virtual cyclic test strain load pseudo After times of damage caused, use the Miner Rule for the time-domain signal ε of the vibration fatigue load of the heat exchanger HP For all the stress change amplitudes and their occurrence times in the rain flow matrix corresponding to ε pseudo The test times are accumulated to obtain the vibration fatigue cycle times with the same fatigue pseudo-damage amplitude of ε pseudo

[0017] Preferably, in the step B), the minimum sampling frequency of the strain gauge and the three-direction vibration acceleration sensor is:

[0018]

[0019] In the formula, N is the maximum engine speed, unit rpm, and c is the number of engine cylinders.

[0020] Preferably, in the step B), define the test cycle according to the defined engine test procedure, continuously test multiple repeated cycle conditions, and the number of test cycles is based on the time-domain change amplitude of each measuring point parameter reaching stability.

[0021] Preferably, in the step C2), the resolution is not less than 100.

[0022] Preferably, in the step C5), the resolution is not less than 100.

[0023] Compared with the prior art, the present invention has the following advantages: It can quickly, efficiently and accurately analyze and evaluate the virtual damage of various fatigue loads of the heat exchanger on the engine test bench and conduct quantitative comparison, clarify the root cause of failure, and help the heat exchanger developers quickly improve the reliable performance of the heat exchanger product. Specific embodiments

[0024] The following further elaborates the present invention in detail with specific embodiments.

[0025] A test method for the failure cause of a heat exchanger based on digital filtering and virtual damage, comprising the following steps:

[0026] ​A) Sensor arrangement: Through finite element analysis of the heat exchanger and combining with the after-sales failure experience of similar products, evaluate the working conditions of the heat exchanger, select the stress concentration areas and failure-prone areas of the heat exchanger as the measuring points to arrange strain gauges, measure the strain signal ε of each measuring point during engine testing, and arrange three-direction vibration acceleration sensors on the heat exchanger shell to measure the acceleration time-domain signal g of the heat exchanger during testing;

[0027] B) Start the engine for testing to obtain the acceleration time-domain signal g and the strain signal ε of each measuring point;

[0028] C) By filtering the strain signal ε of each measuring point during the stable engine test cycle of the heat exchanger and calculating the pseudo-damage, obtain the thermal fatigue cycle number N pseudo corresponding to the standard amplitude ε T of the same virtual cycle test strain load V and the vibration fatigue cycle number N T . Select the maximum value N Tmax of the thermal fatigue cycle number N V and the maximum value N Vmax of the vibration fatigue cycle number N Vmax for numerical comparison. If N Tmax >N Vmax , then vibration fatigue is the main cause of heat exchanger failure during this engine test. Otherwise, thermal fatigue is the main failure cause. The measuring points corresponding to N Tmax are the vibration fatigue weak points and the thermal fatigue weak points respectively.

[0029] Among them, in step C), the thermal fatigue load pseudo-damage calculation includes the following steps:

[0030] C1) Perform Butterworth low-pass filtering on the strain signal ε of each measuring point during the stable engine test cycle of the heat exchanger. The filtering cut-off frequency is 10 hz, and the filtering type is selected to retain the phase. The obtained strain time-domain signal ε LP is used as the thermal fatigue load time-domain signal ε LP of the heat exchanger during the engine test cycle. In this embodiment, a thermocouple can be arranged near the strain gauge measuring point to measure its material temperature. The pulsation amplitude of ε LP has a strong correlation with the change of the material temperature;

[0031] C2) Set the resolution and perform rainflow counting method analysis on the thermal fatigue load time-domain signal ε LP obtained in step C1). Calculate the occurrence times of different amplitude strain changes based on the resolution, and generate the rainflow matrix of the thermal fatigue load time-domain signal ε LP . Assume that the slope of the S-N fatigue life curve of the product is k. During the engine test cycle of the heat exchanger, the amplitude of the thermal fatigue load change is εiLP The occurrence times is n iLP , where i is the resolution and ε pseudo is the standard amplitude of the virtual cyclic test strain load. Then the pseudo-damage of the thermal fatigue load is equivalent to the standard amplitude ε of the virtual cyclic test strain load pseudo After times of damage caused, the Miner Rule is used for the time-domain signal ε of the thermal fatigue load of the heat exchanger LP in the rainflow matrix of all stress change amplitudes and their occurrence times corresponding to ε pseudo of the test times are accumulated to obtain the thermal fatigue cycle times with the same fatigue pseudo-damage amplitude of ε pseudo where j is the upper limit of the resolution.

[0032] In step C), the calculation of the pseudo-damage of the vibration fatigue load includes the following steps:

[0033] C3) Perform Butterworth high-pass filtering on the strain signal ε of each measuring point, with the filtering cut-off frequency of 10 hz and the filtering type selected to retain the phase. The obtained strain time-domain signal ε after filtering HP ;

[0034] C4) Perform fast Fourier transform on the strain time-domain signal ε HP to obtain the vibration strain frequency spectrum diagram, and perform fast Fourier transform on the acceleration time-domain signal g to obtain the acceleration frequency spectrum diagram. Check the change trends and extreme frequencies of the vibration strain frequency spectrum diagram and the acceleration frequency spectrum diagram. If the change trends and extreme frequencies of the two are the same, then the strain time-domain signal ε HP is used as the vibration fatigue load time-domain signal ε of the heat exchanger during the engine test cycle HP , if they are inconsistent, then adjust the filtering cut-off frequency of the high-pass filtering, correct the pressure load of the strain signal ε of the heat exchanger, and analyze and eliminate the noise of the strain signal ε until the change trends and extreme frequencies of the two are the same;

[0035] C5) Set the resolution and perform rainflow counting method analysis on the vibration fatigue load time-domain signal ε obtained in step C4 HP respectively, calculate and count the occurrence times of different amplitude strain changes based on the resolution, and generate the rainflow matrix of the vibration fatigue load time-domain signal ε HP . Let the slope of the S-N fatigue life curve of the product be k, and the change amplitude of the vibration fatigue load of the heat exchanger during the engine test cycle be ε iHP The occurrence times is n iHP , where i is the resolution and ε pseudo is the standard amplitude of the virtual cyclic test strain load. Then the damage of the vibration fatigue load is equivalent to the standard amplitude ε of the virtual cyclic test strain load pseudo After​ For the damage caused each time, the Miner Rule is used to process the time-domain signal ε of the vibration fatigue load of the heat exchanger HP in the rainflow matrix of all stress change amplitudes and their occurrence times corresponding to ε pseudo of the test times are accumulated to obtain the vibration fatigue cycle times with the same fatigue pseudo-damage amplitude of ε pseudo

[0036] In this embodiment, in step B), the minimum sampling frequencies of the strain gauge and the three-direction vibration acceleration sensor are:

[0037]

[0038] In the formula, N is the maximum engine speed, in rpm, and c is the number of engine cylinders.

[0039] In addition, in step B), according to the defined engine test procedure, the test cycle is defined, and multiple repeated cycle conditions are continuously tested. The number of test cycles is based on the time-domain change amplitude of the parameters at each measurement point reaching stability.

[0040] Finally, in this embodiment, in steps C2) and C5), the resolution is not less than 100.

[0041] The test method for the failure cause of the heat exchanger based on digital filtering and virtual damage of the present invention can quickly, efficiently and accurately analyze and evaluate the virtual damage of various fatigue loads of the heat exchanger on the engine test bench and conduct quantitative comparison to clarify the root cause of failure.​

Claims

1. A test method for the failure cause of a heat exchanger based on digital filtering and virtual damage, characterized in that: Including the following steps: A) Sensor arrangement: Through finite element analysis of the heat exchanger and combining with the after-sales failure experience of similar products, evaluate the working conditions of the heat exchanger, select the stress concentration parts and the easily failed parts of the heat exchanger as the measuring points to arrange strain gauges, measure the strain signal ε of each measuring point during the engine test, and arrange three-direction vibration acceleration sensors on the heat exchanger shell to measure the acceleration time-domain signal g of the heat exchanger during the test; B) Start the engine for testing to obtain the acceleration time-domain signal g and the strain signal ε of each measuring point; C) By filtering the strain signal ε of each measuring point of the heat exchanger during a stable engine test cycle and calculating pseudo damage, the standard amplitude ε of the strain load corresponding to the same virtual cycle test is obtained. pseudo The number of thermal fatigue cycles N T and vibration fatigue cycle number N V , select the number of thermal fatigue cycles N T The maximum value N Tmax and vibration fatigue cycle number N V The maximum value N Vmax Perform numerical comparison. If N Vmax >N Tmax , then vibration fatigue is the main reason for the failure of the heat exchanger during this engine test, otherwise thermal fatigue is the main reason for failure, N Vmax and N Tmax The corresponding measuring points are vibration fatigue weak points and thermal fatigue weak points.

2. The test method for the failure cause of a heat exchanger based on digital filtering and virtual damage according to claim 1, characterized in that: In step C), the calculation of the pseudo-damage of the thermal fatigue load includes the following steps: C1) Perform Butterworth low-pass filtering on the strain signal ε at each measurement point of the heat exchanger during a stable engine test cycle. The filtering cut-off frequency is 10 Hz, and the filtering type is selected to retain the phase. The resulting strain time-domain signal ε LP , which serves as the time-domain signal ε of the thermal fatigue load of the heat exchanger during the engine test cycle LP ; C2) Set the resolution for the thermal fatigue load time-domain signal ε obtained in step C1) LP Perform rainflow counting analysis, calculate the occurrence times of different amplitude strain changes based on the resolution, and generate the rainflow matrix of the thermal fatigue load time-domain signal ε LP Let the slope of the product S-N fatigue life curve be k. When the heat exchanger is in the engine test cycle, the amplitude of the thermal fatigue load change is ε iLP The occurrence times is n iLP , i is the resolution, ε pseudo is the standard amplitude of the virtual cycle test strain load. Then the pseudo-damage of the thermal fatigue load is equivalent to the damage caused by the standard amplitude ε pseudo After times, use the Miner Rule to accumulate the test times corresponding to all the stress change amplitudes and their occurrence times in the rainflow matrix of the thermal fatigue load time-domain signal ε LP to obtain the thermal fatigue cycle times with the same fatigue pseudo-damage amplitude of ε pseudo where j is the upper limit of the resolution. pseudo j is the upper limit of the resolution.​ 3. The test method for the failure cause of the heat exchanger based on digital filtering and virtual damage according to claim 1, characterized in that: In step C), the calculation of the pseudo-damage of the vibration fatigue load includes the following steps: C3) Perform Butterworth high-pass filtering on the strain signal ε at each measurement point. The filter cut-off frequency is 10 Hz, and the filter type is selected to preserve the phase. The strain time-domain signal ε obtained after filtering HP ; C4) Strain time-domain signal ε HP Perform a fast Fourier transform on the strain time-domain signal ε to obtain a vibration strain frequency spectrum diagram, and perform a fast Fourier transform on the acceleration time-domain signal g to obtain an acceleration frequency spectrum diagram. Check the change trends and extreme value frequencies of the vibration strain frequency spectrum diagram and the acceleration frequency spectrum diagram. If the change trends and extreme value frequencies of the two are consistent, then use the strain time-domain signal ε HP as the vibration fatigue load time-domain signal ε of the heat exchanger during the engine test cycle HP , if they are inconsistent, then adjust the filter cut-off frequency of the high-pass filter, perform pressure load correction on the strain signal ε of the heat exchanger, and analyze and eliminate the noise of the strain signal ε until the change trends and extreme value frequencies of the two are consistent; C5) Set the resolution for the vibration fatigue load time-domain signal ε obtained in step C4) HP Conduct rainflow counting method analysis respectively, calculate the occurrence times of different amplitude strain changes based on the resolution, and generate the rainflow matrix of the vibration fatigue load time-domain signal ε HP Assume the slope of the S-N fatigue life curve of the product is k, the amplitude change of the vibration fatigue load during the engine test cycle of the heat exchanger is ε iHP The occurrence times is n iHP , i is the resolution, ε pseudo is the standard amplitude of the virtual cyclic test strain load, then the vibration fatigue load damage is equivalent to the damage caused by the standard amplitude ε pseudo after times. Use Miner Rule to accumulate the test times corresponding to all stress change amplitudes and their occurrence times in the rainflow matrix of the vibration fatigue load time-domain signal ε HP to obtain the vibration fatigue cycle times with the same fatigue pseudo-damage amplitude of ε pseudo pseudo ​​ 4. The test method for the failure cause of the heat exchanger based on digital filtering and virtual damage according to claim 1, characterized in that: In step B), the minimum sampling frequency of the strain gauges and the three-direction vibration acceleration sensors is: In the formula, N is the maximum engine speed, in rpm, and c is the number of engine cylinders.

5. The test method for the failure cause of a heat exchanger based on digital filtering and virtual damage according to claim 1, characterized in that: In step B), define the test cycle according to the defined engine test procedure, continuously test multiple repeated cycle conditions, and the number of test cycles is based on the time-domain change amplitude of the parameters of each measuring point reaching stability.

6. The test method for the failure cause of a heat exchanger based on digital filtering and virtual damage according to claim 2, characterized in that: In step C2), the resolution is not less than 100.

7. The method for testing the failure cause of a heat exchanger based on digital filtering and virtual damage according to claim 3, wherein: In step C5), the resolution is not less than 100.

Images