A high temperature resistance test method for data recognition based foamed silicone rubber

Abstract

The present application relates to the technical field of high temperature resistance test of foamed silicone rubber based on data recognition, and discloses a high temperature resistance test method of foamed silicone rubber based on data recognition. The steps include system parameter setting, test device building, short pulse excitation sampling, frequency domain response extraction, temperature step setting, main frequency statistical analysis, normalized frequency shift ratio and second-order difference acceleration calculation, failure determination and retest confirmation. By constructing a complete data-driven test process, the problems of low efficiency, dependence on subjective experience and difficulty in accurately determining the failure point of traditional thermal failure test are solved. The non-destructive spectral analysis method is used to improve the test sensitivity and repeatability, the data-based determination method enhances the objectivity and automation degree, the retest mechanism is introduced to ensure the reliability of the determination, and finally the structured test report is generated.

Description

Technical Field

[0001] This invention relates to the field of high-temperature resistance testing technology for foamed silicone rubber based on data recognition, specifically a high-temperature resistance testing method for foamed silicone rubber based on data recognition. Background Technology

[0002] Foamed silicone rubber is widely used in aerospace, automotive electronics, and high-temperature sealing fields due to its excellent heat resistance, thermal insulation, and mechanical cushioning properties. In these applications, accurately assessing the failure temperature of foamed silicone rubber under high-temperature environments is crucial for material formulation optimization, service life prediction, and reliability design. Existing methods for evaluating high-temperature performance mainly focus on thermogravimetric analysis, differential scanning calorimetry, and traditional dynamic mechanical analysis, with a few based on acoustic emission or infrared thermography.

[0003] Traditional thermogravimetric analysis (TGA) identifies thermal decomposition or weight loss inflection points by measuring the mass-temperature curve of a sample. While it provides a general thermal decomposition temperature, the discreteness of heating rates and mass changes leads to inaccurate failure temperature determination. Furthermore, this method only obtains macroscopic weight loss characteristics and cannot reflect changes in the material's internal microstructure. Differential scanning calorimetry (DSC) focuses on detecting exothermic or endothermic peaks, but for porous, multi-interface systems like foamed silicone rubber, the heat flow signals are often complex, with overlapping peaks and susceptibility to environmental interference, resulting in unstable assessment results. Dynamic mechanical analysis (DMO) measures the elastic modulus and loss modulus of materials at different temperatures, revealing mechanical property degradation characteristics. However, this technique requires complex experimental setups, involves lengthy measurement processes, is sensitive to amplitude control, and demands high uniformity of the sample temperature field, making real-time monitoring of failure points difficult. Recent research has also attempted to utilize acoustic emission (AE) technology to monitor microcrack or pore propagation at high temperatures, indirectly determining failure through spectral analysis of the acoustic emission signals. However, acoustic emission signals are significantly affected by device coupling, external noise, and sample geometry, making it difficult to stably extract spectral characteristics. Furthermore, traditional spectral difference or threshold determination methods rely heavily on empirical threshold settings, lacking a unified and repeatable mathematical model. Other researchers have used infrared imaging to perform time-series analysis of surface temperature fields and radiation intensity to identify thermal runaway or chemical decomposition; however, this method requires extremely high infrared detector resolution and background radiation control, and is significantly affected by device calibration drift, making it difficult to apply in industrial settings. In addition, while existing high-temperature performance evaluation schemes based on resonant frequency detection can capture changes in internal structural stiffness through excitation-response methods, they are mostly single-point or limited-point experiments with large temperature step sizes. Moreover, the extraction and determination of the dominant frequency mainly rely on the absolute value or first derivative change of the maximum amplitude peak, failing to fully utilize the "acceleration" information of the second derivative of frequency shift with respect to temperature. This results in limited resolution for failure temperature location and susceptibility to noise-induced misjudgments. Furthermore, existing technologies lack a mechanism for statistically adaptive threshold calculation of the initial fluctuation range in the data processing stage, often relying on empirical settings or single peak comparisons, making it difficult to balance detection sensitivity and anti-interference capabilities.

[0004] Therefore, this study aims to propose a data-based method for high-temperature resistance testing of foamed silicone rubber. By dynamically analyzing the resonant frequency response of foamed silicone rubber at different temperatures, the method achieves automatic identification of its failure temperature. The method integrates multiple frequency domain analysis techniques, such as normalized frequency shift, second-order differential frequency acceleration, and statistical threshold models. It extracts the dominant frequency variation trend from multi-temperature point measurements, identifies spectral abrupt changes using statistical methods, determines the material failure location, and enhances the reliability of the results through a retesting mechanism. Summary of the Invention

[0005] This invention provides a high-temperature resistance testing method for foamed silicone rubber based on data recognition, which helps to solve the problems mentioned in the background art.

[0006] This invention provides the following technical solution: a high-temperature resistance testing method for foamed silicone rubber based on data recognition, comprising:

[0007] Determine the maximum resonant frequency, sampling frequency, frequency resolution, sampling window length, and number of sampling points, and configure a test device consisting of a piezoelectric transducer, thermocouple, and rigid support.

[0008] The time-domain response signal is acquired by multiple short pulse excitations, the frequency-domain response is calculated and the main resonant frequency is extracted, and the frequency mean and standard deviation are statistically analyzed.

[0009] Set the temperature step size and maximum test temperature, determine the total number of measuring points and gradually increase the temperature to the target temperature at each point, and collect the response data after the temperature stabilizes.

[0010] The excitation and response were repeated at each measurement point, the frequency domain data were calculated, and the mean and variance of the main resonant frequency were statistically analyzed.

[0011] Calculate the normalized frequency shift ratio based on the main frequency variation at each measuring point;

[0012] Based on the frequency response sequence, the second-order differential acceleration of frequency change between adjacent measurement points is calculated;

[0013] Select the non-damage fluctuation range, calculate its second-order difference mean and standard deviation, set the judgment threshold, and identify the failure temperature exceeding the threshold;

[0014] Set up retest temperature points near the identified failure temperature points, resample and recalculate, verify the consistency of the judgment, and generate the final test report.

[0015] Optionally, the step of determining the maximum resonant frequency, sampling frequency, frequency resolution, sampling window length, and number of sampling points, and configuring the test device consisting of a piezoelectric transducer, thermocouple, and rigid support, specifically includes:

[0016] Let the maximum resonant frequency be Therefore, the sampling frequency is selected as ;

[0017] Assume the required frequency resolution is The sampling window length is ;

[0018] Calculate the number of sampling points ;in, It is a rounding function;

[0019] Take a piece of foamed silicone rubber sample, attach a piezoelectric transducer to one end of the sample, and fix the other end to a rigid bracket.

[0020] A thermocouple is attached to the center of the sample to improve temperature measurement accuracy. .

[0021] Optionally, the step of acquiring the time-domain response signal through multiple short-pulse excitations, calculating the frequency-domain response and extracting the main resonant frequency, and statistically analyzing the frequency mean and standard deviation specifically includes:

[0022] at room temperature Next, proceed The short pulse excitation was recorded. Sub-time domain response signal ;

[0023] Do each response Calculate the first Sub-frequency domain response ;

[0024] Extract the first Secondary resonant frequency ;

[0025] Calculate the baseline frequency mean ;

[0026] Calculate the standard deviation of baseline frequencies ;

[0027] Set baseline frequency mean Set the baseline frequency shift ratio .

[0028] Optionally, the setting of the temperature step size and maximum test temperature, determining the total number of measuring points and gradually increasing the temperature to the target temperature, and collecting response data after the temperature stabilizes, specifically includes:

[0029] Set temperature step ;

[0030] Let the maximum test temperature be Total number of measuring points ;

[0031] For the Target temperature at the secondary measurement point To raise the temperature; among them, For measurement point index;

[0032] When the thermocouple readings satisfy At the same time, maintain the sampling window length Response data collection is performed; among which, This refers to the actual measured temperature.

[0033] Optionally, the step of repeatedly exciting and recording the response at each measurement point, calculating the frequency domain data, and statistically analyzing the mean and variance of the main resonant frequency specifically includes:

[0034] For each measuring point Perform the following steps:

[0035] S101, Repeated excitation, record the first... Measurement point number Sub-time domain response signal ;

[0036] S102, Perform each response Calculate the first Measurement point number Sub-frequency domain response :

[0037] ;

[0038] S103, Extract the first Measurement point number Secondary resonant frequency ;

[0039] S104, Calculate the... Mean frequency of measuring points ;

[0040] S105, Calculate the first Measurement point frequency variance .

[0041] Optionally, the step of calculating the normalized frequency shift ratio based on the main frequency variation at each measuring point specifically includes:

[0042] For each Calculate the first Based on the frequency variation at each measuring point, the normalized frequency shift ratio is calculated. .

[0043] Optionally, the step of calculating the second-order differential acceleration of frequency change between adjacent measuring points based on the frequency response sequence specifically includes:

[0044] For each Calculate the first The frequency shift at the measurement point is based on the frequency response sequence, and the second-order differential acceleration of the frequency change between adjacent measurement points is calculated. :

[0045] .

[0046] Optionally, the step of selecting the non-damage fluctuation range, calculating its second-order difference mean and standard deviation, setting a judgment threshold, and identifying the failure temperature exceeding the threshold specifically includes:

[0047] Take the index of the non-damaging fluctuation range , ;in, This serves as the index for the endpoint of the initial statistical interval.

[0048] Calculate the mean of the second difference of the initial interval ;

[0049] Calculate the standard deviation of the second difference of the initial interval ;

[0050] Set the judgment threshold ;

[0051] since The first to meet season And terminate the testing; among them, The identified failure temperature.

[0052] Optionally, the step of setting up retest temperature points near the identified failure temperature points, resampling and calculating, verifying the consistency of the judgment, and generating a final test report specifically includes:

[0053] Obtain the identified failure temperature index Corresponding temperature ;

[0054] Set three points for retesting temperature: ;

[0055] For each temperature point, perform Secondary excitation sampling and calculation and ;

[0056] Calculate the second difference for three points ;

[0057] Verification conditions: , , ;

[0058] If multiple tests are still conducted... If the condition is determined to be invalid, then the temperatures identified in each instance are recorded as a set:

[0059] ;

[0060] Calculate the minimum value With the maximum value :

[0061] , ;

[0062] Let the uncertainty interval be... ;

[0063] Create a table to list each measurement point. of: ;

[0064] Mark failure points with asterisks. .

[0065] The present invention has the following beneficial effects:

[0066] 1. By establishing a reasonable sampling model through the inherent mathematical relationship between parameters such as resonant frequency, sampling frequency, and frequency resolution, a reasonable sampling model is established to ensure that the FFT results have both the required resolution and satisfy the Nyquist sampling condition, thereby improving the accuracy of subsequent signal processing from the source. Simultaneously, by combining the physical layout logic of the piezoelectric transducer and rigid support, the excitation and response paths are ensured to be symmetrical and stable, making the test system both engineering feasible and guaranteeing the consistency of structural response. Compared with existing thermal testing techniques, traditional methods rely on thermal deformation or stress analysis. This scheme characterizes stiffness degradation through frequency variables, avoiding high-power stress loading or irreversible damage, and is particularly suitable for micro-damage monitoring of flexible and foam-like polymer materials.

[0067] 2. A spectral distribution model based on a "room temperature baseline" was established. Steady-state statistics of the material's natural frequencies were achieved through multiple short-pulse excitations, and the mean and standard deviation were used to characterize the range of natural fluctuations, thus providing a "normalized" reference benchmark for judging deviations during subsequent temperature rise. This method combines theoretical support with data stability. Traditional approaches often ignore the spectral fluctuations of materials in their initial state; this scheme incorporates these fluctuations into the statistical model, effectively distinguishing between "natural fluctuations" and abrupt changes caused by "material failure," thereby improving the confidence level of failure identification.

[0068] 3. By controlling the temperature step size, the frequency change acquired at each temperature point is made to be greater than an order of magnitude greater than the standard deviation of the baseline frequency shift, thereby achieving frequency-level distinguishability of temperature changes. This "difference distinguishability control" strategy improves the sensitivity and temperature matching of dynamic spectral analysis. Unlike existing temperature rise loading methods that mostly use 10℃ units, this method can dynamically adjust the step size according to the material response, providing adjustability and analytical flexibility, and better meeting the fine-grained temperature rise resolution requirements in engineering scenarios.

[0069] 4. Performing multiple excitations at each measurement point, recording the response, and conducting frequency domain statistics are key to the noise resistance of this method. Compared to single-shot spectrum acquisition, the calculation of average frequency and variance provides confidence assessment, avoiding the misleading influence of single anomalies on the overall conclusion. Multiple acquisitions form a spectrum cloud map, and the extraction of central features through statistical methods increases the stability of the frequency-shifted signal. Compared to existing methods of "single-point sampling + single-curve inference," this scheme is closer to the integration of signal processing and statistics, significantly improving the reliability of the results.

[0070] 5. The proposed method of calculating the normalized frequency shift ratio using the baseline frequency as a reference represents a significant breakthrough in signal standardization processing. Its significance lies in eliminating the interference of initial frequency differences in samples on experimental results, making the results from different samples and different trials comparable. This frequency domain normalization technique fills the gap in existing polymer thermal testing methods that lack a relative scaling system, and it has strong application value, especially in industrial batch consistency verification scenarios.

[0071] 6. Second-order difference analysis of the frequency shift sequence, using frequency change as an identification indicator, simulates a "frequency shift acceleration" response, enhancing the sensitivity to identifying abrupt changes in material properties. Approximating the mathematical derivative through finite difference analysis captures the slope change after minute fluctuations, providing a deeper structural understanding of the spectral response behavior. Compared to directly observing frequency abrupt changes, second-order difference analysis can capture "acceleration-type changes," suitable for identifying the critical point from slow aging to instantaneous failure in materials, representing a highly promising technical approach for monitoring polymer thermal failure.

[0072] 7. In identifying failure temperatures, the introduction of a judgment model based on the mean + standard deviation × n of the "no-damage range" represents a data-driven statistical innovation. Essentially, it transforms the normal behavior pattern of materials into a judgment standard, moving beyond the threshold of human experience and providing an automated and quantifiable method for failure identification. Compared to the traditional method of "subjective judgment + setting a safe temperature," this solution establishes an objective identification mechanism based on data distribution, offering greater replicability and algorithm embeddability, laying the foundation for intelligent, high-throughput material testing.

[0073] 8. Introducing a retesting mechanism and uncertainty analysis. Identified failure points are retested to determine consistency, avoiding erroneous conclusions due to measurement errors or random disturbances. A failure temperature confidence band is formed by setting upper and lower limit ranges. This data-driven verification strategy enhances the scientific rigor and engineering reliability of the experimental results. Simultaneously, data annotation and asterisk marking are introduced to optimize the report structure, ensuring the final test report is not only clear but also fully auditable, facilitating engineering applications and third-party reproduction. Attached Figure Description

[0074] Figure 1This is a schematic diagram of the process of the present invention. Detailed Implementation

[0075] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0076] Example, refer to Figure 1 A high-temperature resistance test method for foamed silicone rubber based on data recognition, comprising:

[0077] Determine the maximum resonant frequency, sampling frequency, frequency resolution, sampling window length, and number of sampling points, and configure a test device consisting of a piezoelectric transducer, thermocouple, and rigid support.

[0078] The time-domain response signal is acquired by multiple short pulse excitations, the frequency-domain response is calculated and the main resonant frequency is extracted, and the frequency mean and standard deviation are statistically analyzed.

[0079] Set the temperature step size and maximum test temperature, determine the total number of measuring points and gradually increase the temperature to the target temperature at each point, and collect the response data after the temperature stabilizes.

[0080] The excitation and response were repeated at each measurement point, the frequency domain data were calculated, and the mean and variance of the main resonant frequency were statistically analyzed.

[0081] Calculate the normalized frequency shift ratio based on the main frequency variation at each measuring point;

[0082] Based on the frequency response sequence, the second-order differential acceleration of frequency change between adjacent measurement points is calculated;

[0083] Select the non-damage fluctuation range, calculate its second-order difference mean and standard deviation, set the judgment threshold, and identify the failure temperature exceeding the threshold;

[0084] Set up retest temperature points near the identified failure temperature points, resample and recalculate, verify the consistency of the judgment, and generate the final test report.

[0085] A data-based method for testing the high-temperature resistance of foamed silicone rubber is proposed, and a complete test process is constructed, including system parameter setting, response acquisition, frequency domain analysis, normalization processing, acceleration calculation, statistical judgment, and retest confirmation. These steps effectively solve the problems of low test efficiency, strong subjectivity, and difficulty in accurately defining failure points in traditional thermal failure testing. The use of frequency domain signal recognition avoids damage to the material; failure judgment through data-driven methods has the advantages of automation, strong objectivity, and high repeatability; the introduction of a retest mechanism improves the reliability of the results; and a structured test report is generated, facilitating subsequent review or quality tracking management. Overall, a standardized, highly sensitive, and non-destructive thermal stability testing system is constructed, particularly suitable for high-temperature performance research scenarios of flexible and foamed materials.

[0086] The test apparatus, which determines the maximum resonant frequency, sampling frequency, frequency resolution, sampling window length, and number of sampling points, and configures a piezoelectric transducer, thermocouple, and rigid support, specifically includes:

[0087] Let the maximum resonant frequency be To determine the upper limit of the highest possible resonant frequency of the sample, in order to guide the selection of the subsequent sampling frequency; therefore, the sampling frequency is selected as... It satisfies the Nyquist sampling theorem, avoiding spectral aliasing;

[0088] Assume the required frequency resolution is Define the minimum identifiable frequency interval required for frequency domain analysis; then the sampling window length is... Ensure that the FFT resolution meets the requirements. ;

[0089] Calculate the number of sampling points ;in, This is the floor function; calculate the required number of sampling points.

[0090] Take a piece of foamed silicone rubber sample, attach a piezoelectric transducer to one end of the sample, and fix the other end to a rigid support; ensure that the excitation and response are coupled in a consistent manner.

[0091] A thermocouple is attached to the center of the sample to improve temperature measurement accuracy. High-precision recording of sample temperature ensures stable temperature conditions.

[0092] The key is to clearly define the frequency parameter system (maximum resonant frequency, sampling frequency, number of sampling points, etc.) and physical test device configuration (piezoelectric transducer, thermocouple, rigid support, etc.) upon which the test relies. This step effectively solves problems such as spectral aliasing, insufficient sampling, or inaccurate temperature monitoring encountered in previous tests. Specifically, the sampling frequency is set according to the Nyquist criterion based on the maximum resonant frequency to avoid frequency folding; the number of sampling points ensures that the FFT has the required frequency resolution, improving the accuracy of spectral analysis; the piezoelectric transducer and rigid support structure ensures clear excitation and response paths; and the thermocouple mounting position is designed reasonably to ensure that the temperature measurement accurately reflects the core temperature of the sample. A set of test hardware and parameter design specifications with a rigorous theoretical foundation is provided to ensure the accuracy, consistency, and physical repeatability of subsequent experimental data, providing a fundamental guarantee for the entire frequency domain failure identification system.

[0093] The process of acquiring time-domain response signals through multiple short-pulse excitations, calculating frequency-domain response and extracting the main resonant frequency, and statistically analyzing the frequency mean and standard deviation specifically includes:

[0094] at room temperature Next, proceed The short pulse excitation was recorded. Sub-time domain response signal Collect sufficient samples and statistically analyze the baseline frequency distribution;

[0095] Do each response Calculate the first Sub-frequency domain response Convert the time-domain signal to the frequency domain;

[0096] Extract the first Secondary resonant frequency The main resonant frequency corresponding to the peak value of the positioning spectrum;

[0097] Calculate the baseline frequency mean ;

[0098] Calculate the standard deviation of baseline frequencies ;

[0099] The mean and standard deviation of the baseline frequency are calculated to provide a basis for normalization and threshold setting;

[0100] Set baseline frequency mean Set the baseline frequency shift ratio Determine the normalized reference frequency and the initial ratio.

[0101] The frequency response at room temperature was acquired through multiple short-pulse excitations, and the mean and standard deviation of the dominant frequency were statistically analyzed to establish a "baseline spectrum model" for the material. This step addresses the problem in traditional methods where the lack of an initial benchmark leads to the inability to quantify subsequent temperature rise spectrum changes. By statistically analyzing the frequency mean and standard deviation of a large number of samples, a reference system for normalization and a noise boundary for the judgment model were constructed, providing a quantitative basis for subsequent frequency shifts and abnormal changes. This improves the sensitivity of frequency shift detection, enabling the differentiation of minute thermal effect changes; provides a basis for error tolerance, enhancing the stability and anti-interference capability of the judgment model; and provides fundamental data support for failure threshold calculation, making the entire data processing process structured and mathematically supported.

[0102] The process involves setting the temperature step size and maximum test temperature, determining the total number of measuring points, and gradually increasing the temperature at each point to the target temperature. After the temperature stabilizes, response data is collected. Specifically, this includes:

[0103] Set temperature step Ensure that each frequency shift exceeds one-tenth of the baseline standard deviation and can be reliably resolved;

[0104] Let the maximum test temperature be Total number of measuring points The number of measurement points required to cover the target temperature range;

[0105] For the Target temperature at the secondary measurement point To raise the temperature; among them, For measurement point index;

[0106] When the thermocouple readings satisfy At the same time, maintain the sampling window length Response data collection is performed; among which, To ensure the actual measured temperature; data is collected under stable target temperature conditions to eliminate the effects of temperature drift.

[0107] By setting reasonable temperature step sizes and target maximum temperatures, distributed acquisition of response data solves the problems of uneven distribution of spectrum sampling points and significant interference from temperature fluctuations in traditional continuous heating methods. In particular, setting the step size to 1 / 10 of the standard deviation of the baseline frequency as a discrimination requirement ensures that frequency changes caused by each temperature rise step can be significantly detected, thereby improving analytical resolution and stability. Enhanced controllability of the spectrum response under temperature control results in a clear and monotonic frequency shift trend; uniform coverage of measurement points in the temperature space is achieved, facilitating the construction of a complete temperature-frequency response curve; and data acquisition is ensured only after the target temperature has stabilized, effectively eliminating transient interference caused by temperature disturbances and guaranteeing signal quality.

[0108] The process of repeatedly exciting and recording the response at each measurement point, calculating frequency domain data, and statistically analyzing the mean and variance of the main resonant frequency specifically includes:

[0109] For each measuring point Perform the following steps:

[0110] S101, Repeated excitation, record the first... Measurement point number Sub-time domain response signal ; Obtain multiple response samples at the current temperature;

[0111] S102, Perform each response Calculate the first Measurement point number Sub-frequency domain response :

[0112] Converting to the frequency domain facilitates frequency extraction.

[0113] S103, Extract the first Measurement point number Secondary resonant frequency Identify the main resonant frequency;

[0114] S104, Calculate the... Mean frequency of measuring points ;

[0115] S105, Calculate the first Measurement point frequency variance ;

[0116] Calculate the mean and variance of the current temperature frequency to assess measurement consistency.

[0117] Multiple excitation and frequency domain response calculations are performed at each temperature point, and the frequency mean and variance are statistically analyzed to enhance data stability. This step effectively solves the problem of "single acquisition being greatly affected by random noise, resulting in unstable results" in traditional methods. By repeatedly sampling to form a frequency distribution cloud map, stable feature values ​​are extracted, thus avoiding the risk of accidental disturbances misleading failure judgments. This improves the accuracy of frequency change trend judgment; provides high-quality input data for subsequent normalization and acceleration calculations; and through the joint control of mean and variance, it has the ability to dynamically identify data outliers, providing a good foundation for identifying the nonlinear response of complex materials.

[0118] The calculation of the normalized frequency shift ratio based on the main frequency variation at each measuring point specifically includes:

[0119] For each Calculate the first Based on the frequency variation at each measuring point, the normalized frequency shift ratio is calculated. Eliminate absolute frequency differences and quantize frequency shifts into relative ratios.

[0120] By calculating the normalized frequency shift ratio at each measurement point, the original frequency change is standardized, resolving the interference of fundamental frequency differences between different samples or batches on the overall judgment result. This improves the consistency and comparability of cross-sample comparisons, facilitating standardized judgments in production or quality control processes. Normalization effectively reduces the deviation of "absolute frequency" caused by non-temperature factors such as structural dimensions and boundary conditions, making frequency shift changes more focused on the causal relationship between "temperature and performance." It also provides support for unifying the scale and simplifying the model structure of subsequent differential and failure determination algorithms, improving the versatility and mathematical interpretability of the data model.

[0121] The calculation of the second-order differential acceleration of frequency change between adjacent measurement points based on the frequency response sequence specifically includes:

[0122] For each Calculate the first The frequency shift at the measurement point is based on the frequency response sequence, and the second-order differential acceleration of the frequency change between adjacent measurement points is calculated. :

[0123] The abrupt change in frequency shift acceleration is reflected by approximating the second derivative of the frequency shift with respect to temperature using finite difference.

[0124] This method introduces the second-order difference acceleration of the frequency shift sequence as an identification variable, essentially performing finite-difference calculations on the second derivative of the normalized frequency shift ratio. This approach breaks through the "linear assumption" of a single frequency change, enabling more sensitive identification of abrupt changes, inflection points, and nonlinear variation characteristics, thus addressing the shortcomings of existing methods in identifying sudden material failure responses. By measuring acceleration, the "abrupt trend" of the frequency response can be accurately detected as a signal of the material's transition from steady state to instability. Acceleration, as a higher-order statistical feature, is more suitable for clustering multiple samples or identifying anomalies, improving the recognition rate and robustness of the intelligent judgment model. It also provides more significant discriminative features for subsequent statistical threshold determination, resulting in a significant reduction in the false alarm rate.

[0125] The process of selecting a non-damaging fluctuation range, calculating its second-order difference mean and standard deviation, setting a judgment threshold, and identifying failure temperatures exceeding the threshold specifically includes:

[0126] Take the index of the non-damaging fluctuation range , ;in, Used as the initial statistical interval endpoint index; This determines the non-damaging fluctuation interval for statistical threshold calculation.

[0127] Calculate the mean of the second difference of the initial interval ;

[0128] Calculate the standard deviation of the second difference of the initial interval ;

[0129] Set the judgment threshold ;

[0130] Based on the initial interval statistical acceleration distribution, three times the standard deviation is set as the failure judgment threshold;

[0131] since The first to meet season And terminate the testing; among them, The failure temperature is identified; the material failure temperature is identified at the acceleration abrupt change point.

[0132] A statistical standard for a "damage-free fluctuation range" was established, and the mean and standard deviation of the second-order acceleration were calculated based on this standard. This was used to set a dynamic threshold and identify the failure temperature corresponding to abrupt changes in the spectral density. This threshold strategy based on sample distribution solves the problems of poor adaptability and low sensitivity inherent in traditional "manually set fixed thresholds." Dynamic judgment rules are constructed based on real response data, automatically adapting to different materials or test conditions. The statistical discrimination strategy makes the failure point identification process more objective and robust, making it particularly suitable for embedded applications in automated systems. The "statistical confidence interval" set by three times the standard deviation effectively controls the risk of misjudgment, improving specificity while maintaining sensitivity.

[0133] The process of setting up retest temperature points near the identified failure temperature points, resampling and calculating, verifying the consistency of the judgment, and generating a final test report specifically includes:

[0134] Obtain the identified failure temperature index Corresponding temperature ;

[0135] Set three points for retesting temperature: ;

[0136] For each temperature point, perform Secondary excitation sampling and calculation and ;

[0137] Validation only The failure signal appeared at the location, rather than due to misjudgment caused by measurement fluctuations;

[0138] Calculate the second difference for three points ;

[0139] Verification conditions: , , ;

[0140] Accurately eliminate measurement noise and identify only a single temperature point where acceleration abruptly changes, confirming that this point as the true failure temperature.

[0141] If multiple tests are still conducted... If the condition is determined to be invalid, then the temperatures identified in each instance are recorded as a set:

[0142] ;

[0143] Calculate the minimum value With the maximum value :

[0144] , ;

[0145] Let the uncertainty interval be... ;

[0146] Quantitative retesting fluctuations lead to uncertainty in failure temperature;

[0147] Create a table to list each measurement point. of: ;

[0148] Mark failure points with asterisks. ;

[0149] All measurement data are presented in full for easy review and archiving.

[0150] This paper proposes a method to remeasure the temperature at three points for identified failure points and verify the consistency of failure judgment through resampling, while outputting the uncertainty range of the failure temperature. This mechanism solves the problems of fuzzy boundary point judgment and poor repeatability in traditional methods. The retesting mechanism eliminates misjudgments caused by "random anomalies" or measurement noise, improving the reliability of the results. The "uncertainty range" formed by the upper and lower limit values ​​provides more valuable data references for applications such as product quality analysis and process control. The final data report presents information such as frequency shift, differential, and failure markers for each measurement point in a structured manner, making the results easy to track, trace, reproduce, and compare, and possessing good industrialization and engineering adaptability.

[0151] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0152] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for high temperature resistance test of data recognition based foamed silicone rubber, characterized in that, The method comprises the following steps: Determine the maximum resonance frequency, sampling frequency, frequency resolution, sampling time window length and sampling point number, and configure the test device composed of a piezoelectric transducer, a thermocouple and a rigid support; Collect time domain response signals through multiple short pulse excitations, calculate frequency domain responses, extract the main resonance frequency, and calculate the mean and standard deviation of the frequency; Set the temperature step and the maximum test temperature, determine the total number of measurement points, and gradually increase the temperature to the target temperature. After the temperature is stable, collect the response; Repeat the excitation and record the response at each measurement point, calculate the frequency domain data, and calculate the mean and variance of the main resonance frequency; Calculate the normalized frequency shift ratio according to the variation of the main frequency of each measurement point; Calculate the second-order difference acceleration of the frequency variation between adjacent measurement points based on the frequency response sequence; Select a non-damage fluctuation interval, calculate the second-order difference mean and standard deviation, set a judgment threshold, and identify the failure temperature that exceeds the threshold; Set retest temperature points near the identified failure temperature points, re-sample and calculate, and verify the consistency of the judgment to generate a final test report.

2. A method for testing the high temperature resistance of a data-identified foamed silicone rubber according to claim 1, characterized in that, The method for determining the maximum resonance frequency, sampling frequency, frequency resolution, sampling time window length and sampling point number, and configuring the test device composed of a piezoelectric transducer, a thermocouple and a rigid support, specifically comprises: Let the maximum resonant frequency be Thus, take the sampling frequency to be ; Let the required frequency resolution be , then the sampling time window length is ; Computing the number of samples ; wherein is a ceiling function; Take a piece of foamed silicone rubber sample, adhere a piezoelectric transducer to one end of the sample, and fix the other end to a rigid support; A thermocouple is attached at the center of the sample, and the temperature measurement accuracy .

3. A method of testing the high temperature resistance of a data-identified foamed silicone rubber according to claim 2, characterized in that, The method for collecting time domain response signals through multiple short pulse excitations, calculating frequency domain responses, extracting the main resonance frequency, and calculating the mean and standard deviation of the frequency, specifically comprises: at room temperature The following, is performed short-pulse excitation, the first time-domain response signal ; Each time a response is made , the frequency domain response is calculated ; extracting the first secondary resonant frequency ; Calculate mean of baseline frequencies ; Computing baseline frequency standard deviation ; setting a baseline frequency mean , setting a baseline frequency shift ratio .

4. The method for high temperature resistance test of data recognition based foamed silicone rubber according to claim 3, characterized in that, The method for setting the temperature step and the maximum test temperature, determining the total number of measurement points, and gradually increasing the temperature to the target temperature, and collecting the response after the temperature is stable, specifically comprises: Set temperature step ; Let the maximum test temperature be , the total number of measuring points ; for the i-th measurement point target temperature the i-th measurement point target temperature temperature is raised; wherein, is an index of the measurement point When the thermocouple reading meets the sampling time window length is kept response acquisition is carried out; wherein, is the actual measured temperature.

5. A method of testing the high temperature resistance of a data-identified foamed silicone rubber according to claim 4, characterized in that, The method for repeating the excitation and recording the response at each measurement point, calculating the frequency domain data, and calculating the mean and variance of the main resonance frequency, specifically comprises: For each measurement point The following steps are performed: S101, repeat excitation, record the first measurement point the time domain response signal ; S102, Perform each response Calculate the first Measurement point number Sub-frequency domain response : ； S103、extracting the first measurement point the first second main resonance frequency ; S104、calculating the first measurement point frequency mean ; S105、calculating the first measurement point frequency variance .

6. A method of testing the high temperature resistance of a data-identified foamed silicone rubber according to claim 5, characterized in that, The method for calculating the normalized frequency shift ratio according to the variation of the main frequency of each measurement point, specifically comprises: For each , the first measurement point according to the change of the main frequency of each measurement point, the normalized frequency shift ratio is calculated.

7. A method of testing the high temperature resistance of a data-identified foamed silicone rubber according to claim 6, characterized in that, The method for calculating the second-order difference acceleration of the frequency variation between adjacent measurement points based on the frequency response sequence, specifically comprises: For each , the first derivative of the frequency response is calculated : 。 8. A method of testing the high temperature resistance of a data-identified foamed silicone rubber according to claim 7, characterized in that, The method for selecting a non-damage fluctuation interval, calculating the second-order difference mean and standard deviation, setting a judgment threshold, and identifying the failure temperature that exceeds the threshold, specifically comprises: Taking an undamaged fluctuation interval index , ; wherein is an initial statistical interval end index; calculating an initial interval second order difference mean ; calculating an initial interval second order difference standard deviation ; Setting a decision threshold ; Since the first time that is met, the and the detection is terminated; wherein is the identified failure temperature.

9. The method of claim 8, wherein the data-identified foamable silicone rubber is subjected to a high temperature test. The method for setting retest temperature points near the identified failure temperature points, re-sampling and calculating, and verifying the consistency of the judgment to generate a final test report, specifically comprises: acquiring the identified failure temperature index corresponding temperature ; Set retest temperature three points: ; For each temperature point, the following is done Sub-excitation sampling and calculation and ; Second order difference for three-point calculation ; Verification condition: , , ; If the measurement is still multiple times If the measurement is still multiple times ； Computing the minimum value With the maximum value : ， ； Let the uncertainty interval be ; A table is constructed listing each measurement point of: ; Marking the failure point with an asterisk .

Images