Radiator salt spray corrosion life prediction method based on dynamic time warping

Abstract

The invention discloses a radiator salt spray corrosion life prediction method based on dynamic time warping, which belongs to the technical field of material corrosion test and life prediction, and comprises the following steps: collecting time sequence data of surface impedance and thermal resistance change of a radiator, constructing a dual-channel corrosion characteristic data set, and recording a salt spray concentration data value. The temperature data value and the humidity data value are combined, and an environment sensitivity weight vector is established. Accelerated corrosion in a salt spray environment is simulated, a laboratory accelerated corrosion test spectrum is generated, meanwhile, an actual environment fluctuation spectrum is monitored, the two spectrums are aligned by using a weight vector, accelerated corrosion data are obtained, and a basic life prediction value is calculated. And calculating an environment-structure coupling factor by combining the interaction between the three-dimensional structure characteristic parameters of the radiator and the temperature and humidity data values, correcting the basic life prediction value, and generating a final life prediction result. According to the method, double-channel feature alignment and coupling factor correction are adopted, and the corrosion life of the radiator of the specific structure under the actual working condition can be predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of material corrosion testing and life prediction, and in particular to a method for predicting the salt spray corrosion life of a radiator based on dynamic time warping. Background Art

[0002] Radiators, as key heat exchange components, are widely used in automobiles, electronic equipment, and industrial systems. Their long-term stable operation is crucial to the reliability of the entire system. Radiators are typically made of metal materials such as aluminum alloys. During service, especially in marine, coastal, or winter salt-sprayed road environments, they are exposed to humid, salty atmospheres for long periods of time, making them susceptible to electrochemical corrosion. Corrosion not only weakens the material's mechanical strength but also causes the accumulation of corrosion products, blocking heat dissipation channels and reducing heat transfer efficiency, ultimately leading to radiator failure. Therefore, accurately predicting the corrosion life of radiators in salt spray environments is of great engineering value.

[0003] Existing methods for predicting the corrosion life of radiators primarily include traditional salt spray testing and statistical methods based on empirical models. Salt spray testing involves subjecting the radiator to continuous or cyclic accelerated corrosion testing in a standardized salt spray chamber, assessing the lifespan based on the time it takes to reach a specified corrosion level or performance degradation threshold. Statistical methods typically rely on collecting historical failure data from a large number of similar products in different regions, and using regression analysis and other methods to establish empirical formulas linking lifespan to service time or certain macro-environmental parameters.

[0004] However, the above-mentioned existing technical means have obvious limitations. The corrosion stress imposed by traditional accelerated salt spray tests is usually constant and severe, which is very different from the complex environment in which the temperature, humidity, and salt spray concentration of the radiator fluctuate dynamically during actual service. As a result, the correlation between the test results and the actual service life is poor, and the acceleration factor is difficult to determine. The statistical method based on empirical models is heavily dependent on the quantity and quality of historical data, lacking predictive ability for radiators designed with new materials or new structures. Moreover, its model often ignores the physical and chemical mechanisms of the corrosion process and the influence of structural details on corrosion sensitivity, resulting in low prediction accuracy and universality. Summary of the Invention

[0005] To solve the above problems, the present invention provides a radiator salt spray corrosion life prediction method based on dynamic time warping. It adopts a method combining dynamic time warping alignment of dual-channel characteristics with environment-structure coupling factor correction to predict the corrosion life of a radiator with a specific structure under actual working conditions.

[0006] The above objectives can be achieved through the following solutions: The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping includes obtaining the surface impedance time series data and thermal resistance change time series data of the radiator in a salt spray environment to generate a dual-channel corrosion feature data set; deploying a multi-sensor network in the actual use environment of the radiator to record salt spray concentration data values, temperature data values ​​and humidity data values; establishing an environmental sensitivity weight vector based on the dual-channel corrosion feature data set, the salt spray concentration data values, the temperature data values ​​and the humidity data values; simulating accelerated corrosion in the salt spray environment based on the temperature data values ​​and the humidity data values ​​to generate a laboratory accelerated corrosion test spectrum, and through long-term monitoring The radiator collects the actual environmental fluctuation spectrum; based on the environmental sensitivity weight vector, the laboratory accelerated corrosion test spectrum is matched and aligned with the time period with equivalent corrosion effect in the actual environmental fluctuation spectrum to generate time-domain aligned accelerated corrosion data; calculations are performed based on the time-domain aligned accelerated corrosion data to generate a basic life prediction value; the three-dimensional structural characteristic parameters of the radiator are obtained, and the environment-structure coupling factor is calculated by combining the interaction between the three-dimensional structural characteristic parameters and the temperature data value and the humidity data value; the basic life prediction value is corrected by the environment-structure coupling factor to generate a final life prediction result.

[0007] Optionally, generating a dual-channel corrosion feature data set includes: collecting continuous measurement values ​​of the electrochemical impedance spectrum of the radiator surface over time to generate a surface impedance time series graph; synchronously monitoring the attenuation rate of the heat conduction performance of the radiator to generate a thermal resistance change rate sequence; performing timestamp synchronization fusion processing on the surface impedance time series graph and the thermal resistance change rate sequence to generate a dual-channel corrosion feature data set.

[0008] Optionally, establishing the environmental sensitivity weight vector includes: normalizing the salt spray concentration data value, the temperature data value, and the humidity data value to generate a standardized environmental parameter vector; analyzing the correlation between the standardized environmental parameter vector and the dual-channel corrosion feature data set through the Pearson correlation coefficient to determine the sensitivity coefficient of each environmental parameter; integrating the weight relationship between the salt spray concentration data value, the temperature data value, and the humidity data value based on the sensitivity coefficient and performing quantization processing to establish the environmental sensitivity weight vector.

[0009] Optionally, generating a laboratory accelerated corrosion test spectrum and collecting an actual environmental fluctuation spectrum through long-term monitoring of the radiator includes: setting gradient-changing temperature data values ​​and humidity data values ​​in the controllable salt spray environment to simulate the corrosion process under different environmental stress combinations, and generating a laboratory accelerated corrosion test spectrum with a corrosion rate change trend; analyzing the recorded fluctuations of the salt spray concentration data values, the temperature data values, and the humidity data values ​​to generate an actual environmental fluctuation spectrum; and performing time scale normalization processing on the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum.

[0010] Optionally, the generation of time-domain aligned accelerated corrosion data includes: segmenting the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum into time series to generate multiple local time windows; under the constraint of the environmental sensitivity weight vector, calculating the similarity measure value between the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum in each local time window; dynamically adjusting the scaling ratio of the time series according to the similarity measure value to keep the corrosion rate change trend consistent with the actual environmental fluctuation spectrum; and interpolating and reconstructing the laboratory accelerated corrosion test spectrum based on the adjusted time series to generate time-domain aligned accelerated corrosion data.

[0011] Optionally, generating a basic life prediction value includes: calculating the corrosion depth of the time-domain aligned accelerated corrosion data to generate a cumulative corrosion amount curve; determining an initial life node based on the intersection of a preset corrosion amount threshold and the cumulative corrosion amount curve; and linearly compensating the initial life node in combination with the temperature data value and the humidity data value to generate a basic life prediction value.

[0012] Optionally, generating the environment-structure coupling factor includes: performing feature extraction based on the three-dimensional structural characteristic parameters to obtain surface curvature distribution data and weld density distribution data; performing sensitivity analysis on the surface curvature distribution data according to the temperature data value and the humidity data value to generate a temperature-curvature correction coefficient and a humidity-curvature correction coefficient; generating a thermal stress concentration factor based on the interaction calculation between the weld density distribution data and the temperature data value; and combining the temperature-curvature correction coefficient, the humidity-curvature correction coefficient and the thermal stress concentration factor to construct an environment-structure coupling factor.

[0013] Optionally, generating the final life prediction result includes: performing correlation mapping based on the environment-structure coupling factor to generate a life compensation coefficient; performing nonlinear correction on the basic life prediction value according to the life compensation coefficient to generate a corrected life prediction value; and performing smoothing processing based on the corrected life prediction value to generate the final life prediction result.

[0014] Optionally, the method further includes: obtaining a status record of the radiator when it actually fails, and generating actual failure data by combining the surface impedance time series data and the thermal resistance change time series data; comparing the final life prediction result with the actual failure data to generate a prediction error distribution map; reversely adjusting the environmental sensitivity weight vector according to the prediction error distribution map, and updating the calculation rules of the environment-structure coupling factor.

[0015] Based on the same inventive concept, the present invention also provides a radiator salt spray corrosion life prediction system based on dynamic time warping, the system comprising: a dual-channel data module for acquiring the surface impedance time series data and thermal resistance change time series data of the radiator in a salt spray environment, and generating a dual-channel corrosion feature data set; an environmental detection module for deploying a multi-sensor network in the actual use environment of the radiator, and recording salt spray concentration data values, temperature data values ​​and humidity data values; a weight vector construction module for establishing an environmental sensitivity weight vector based on the dual-channel corrosion feature data set, the salt spray concentration data values, the temperature data values ​​and the humidity data values; a corrosion simulation generation module for simulating accelerated corrosion in the salt spray environment based on the temperature data values ​​and the humidity data values, and generating a laboratory accelerated corrosion test. The invention discloses a method for detecting the corrosion of a heat sink by using a three-dimensional structure and a structure coupling calculation module, wherein the three-dimensional structure is detected by the heat sink, and the actual environmental fluctuation spectrum is collected by long-term monitoring of the radiator; a time domain alignment processing module is used to match and align the laboratory accelerated corrosion test spectrum with the time period with equivalent corrosion effect in the actual environmental fluctuation spectrum based on the environmental sensitivity weight vector, and generate time domain aligned accelerated corrosion data; a life basic prediction module is used to perform calculations based on the time domain aligned accelerated corrosion data to generate a basic life prediction value; a structure coupling calculation module is used to obtain the three-dimensional structural characteristic parameters of the radiator, and calculate the environment-structure coupling factor by combining the interaction between the three-dimensional structural characteristic parameters and the temperature data value and the humidity data value; a life dynamic correction module is used to correct the basic life prediction value by the environment-structure coupling factor to generate a final life prediction result.

[0016] Compared with the prior art, the present invention has the following advantages: 1. The present invention establishes a dual-channel corrosion feature data set, combining surface impedance data reflecting the microscopic corrosion mechanism with thermal resistance change data characterizing macroscopic performance attenuation, thereby achieving a comprehensive and multi-dimensional characterization of the radiator corrosion degradation process, overcoming the defect of one-sided information of a single monitoring indicator, and providing a more reliable data basis for life prediction.

[0017] 2. The present invention innovatively introduces a physically significant environmental sensitivity weight vector to constrain the dynamic time warping algorithm, aligning the laboratory accelerated corrosion spectrum with the actual environmental fluctuation spectrum, solving the problem of unequal time scales and poor correlation between traditional accelerated tests and actual working conditions, and greatly improving the accuracy of extrapolating laboratory data to actual service life.

[0018] 3. This invention quantifies the interaction between the three-dimensional geometric characteristics of the radiator, such as surface curvature and weld distribution, and environmental factors by constructing an environment-structure coupling factor. This allows the life prediction to reflect the individual life differences of specific products caused by structural design differences, elevating the prediction from a universal material assessment to a precise assessment of specific structural components, and the prediction results are closer to engineering practice.

[0019] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 It is a flow chart of a method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to an embodiment of the present invention.

[0022] Figure 2 1 is a comparison chart of the laboratory and actual environment spectra of an embodiment of the present invention.

[0023] Figure 3 1 is a cumulative corrosion amount curve and a threshold value diagram of an embodiment of the present invention.

[0024] Figure 4 It is a structural schematic diagram of a radiator salt spray corrosion life prediction system based on dynamic time warping according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0026] Reference Figure 1 An embodiment of the present invention proposes a method for predicting the salt spray corrosion life of a radiator based on dynamic time warping. By combining the dynamic time warping alignment of dual-channel features with the correction of the environment-structure coupling factor, the corrosion life of a radiator with a specific structure under actual working conditions can be predicted.

[0027] The method of this embodiment specifically includes: Obtain the surface impedance time series data and thermal resistance change time series data of the radiator in a salt spray environment to generate a dual-channel corrosion feature data set; Deploy a multi-sensor network in the actual use environment of the radiator to record salt spray concentration data values, temperature data values, and humidity data values; Establishing an environmental sensitivity weight vector based on the dual-channel corrosion feature data set, the salt spray concentration data value, the temperature data value, and the humidity data value; Simulating accelerated corrosion in the salt spray environment based on the temperature data value and the humidity data value to generate a laboratory accelerated corrosion test spectrum, and collecting an actual environmental fluctuation spectrum by long-term monitoring of the radiator; Matching and aligning the laboratory accelerated corrosion test spectrum with time periods having equivalent corrosion effects in the actual environmental fluctuation spectrum based on the environmental sensitivity weight vector to generate time-domain aligned accelerated corrosion data; Perform calculations based on the time-domain aligned accelerated corrosion data to generate a basic life prediction value; Acquire three-dimensional structural characteristic parameters of the radiator, and calculate and generate an environment-structure coupling factor by combining the interaction between the three-dimensional structural characteristic parameters and the temperature data value and the humidity data value; The basic life prediction value is corrected by the environment-structure coupling factor to generate a final life prediction result.

[0028] Specifically, the process begins by acquiring time-series data on the surface impedance and thermal resistance of a heat sink in a salt spray environment. These data are then integrated to generate a dual-channel corrosion signature dataset. A multi-sensor network is then deployed in the actual environment. This network consists of at least a high-precision temperature sensor, a humidity sensor, and a salt spray concentration sensor for measuring airborne chloride ion concentration or surface salt deposition rate. These sensors are connected via a data acquisition unit (DAQ) to continuously record and store real-world data on salt spray concentration, temperature, and humidity over time at a preset sampling frequency, such as every minute or every hour. Based on the dual-channel corrosion signature dataset and the environmental data collected by the sensors, an environmental sensitivity weight vector is established to quantify the contribution of different environmental factors to the corrosion process. Next, the temperature and humidity data are used to simulate accelerated corrosion in a salt spray environment to generate a laboratory accelerated corrosion test spectrum. Simultaneously, the actual environmental fluctuation spectrum is collected through long-term monitoring of the heat sink. The laboratory accelerated corrosion test spectrum is aligned with the time periods in the actual environmental fluctuation spectrum that exhibit equivalent corrosion effects using the environmental sensitivity weight vector to generate time-domain aligned accelerated corrosion data. A basic life prediction value is then calculated based on this time-domain aligned accelerated corrosion data. The radiator's three-dimensional structural characteristic parameters are further obtained, and the environment-structure coupling factor is calculated by combining the interaction of temperature and humidity data. This factor is used to modify the basic life prediction value, ultimately obtaining a more accurate life prediction result. This method significantly improves the accuracy of life prediction through multi-dimensional data fusion and dynamic correction. Through dynamic alignment and coupling correction, the actual life of the radiator in complex environments can be more realistically reflected, providing a scientific basis for radiator maintenance and replacement.

[0029] Optionally, generating a dual-channel corrosion feature dataset includes: Collect continuous measurement values ​​of the electrochemical impedance spectrum of the radiator surface over time to generate a surface impedance time series graph; Simultaneously monitor the heat conduction performance decay rate of the radiator and generate a thermal resistance change rate sequence; The surface impedance time series graph and the thermal resistance change rate sequence are subjected to time stamp synchronization fusion processing to generate a dual-channel corrosion feature data set.

[0030] Specifically, an electrochemical workstation is used, with the surface of the radiator as the working electrode, and a reference electrode and a counter electrode form a three-electrode system. During the salt spray test, a small AC voltage is regularly applied to the three-electrode system, and the electrochemical impedance spectrum (such as the Nyquist plot or Bode plot) is measured and arranged in chronological order to form a surface impedance time series diagram. While performing the electrochemical measurement, the instantaneous thermal resistance of the radiator is calculated by applying a stable thermal power to one end of the radiator and using a high-precision temperature sensor array to monitor the temperature difference between the hot end and the cold end in real time. For calculating the thermal resistance of the radiator ,have: , in, is the thermal resistance of the heat sink; is the steady-state temperature difference between the hot end and the cold end of the radiator. This data is directly collected by the deployed temperature sensor. is the stable thermal power applied to the heat sink, and this value is determined by the power of the experimentally controlled heating source. The thermal resistance change rate is obtained by calculating the thermal resistance increment between two adjacent measurement time points, thus forming a time series, namely the thermal resistance change rate series. A synchronized timestamp accurate to the second or millisecond level is marked for each electrochemical impedance spectroscopy measurement and thermal resistance measurement behavior. Based on this timestamp, the key features in the surface impedance time series map obtained at each time point, such as the charge transfer resistance value, are paired with the corresponding thermal resistance change rate value and fused into a structured dual-channel corrosion feature data set. By constructing a dual-channel corrosion feature data set, a richer and more reliable feature foundation is provided for the subsequent use of dynamic time warping to align laboratory accelerated data with actual environment data.

[0031] Optionally, establishing the environmental sensitivity weight vector includes: Normalizing the salt spray concentration data value, the temperature data value, and the humidity data value to generate a standardized environmental parameter vector; Analyzing the correlation between the standardized environmental parameter vector and the dual-channel corrosion characteristic data set through the Pearson correlation coefficient to determine the sensitivity coefficient of each environmental parameter; The weight relationship among the salt fog concentration data value, the temperature data value, and the humidity data value is integrated based on the sensitivity coefficient and quantized to establish an environmental sensitivity weight vector.

[0032] Specifically, the salt spray concentration data values, temperature data values, and humidity data values ​​continuously recorded by a multi-sensor network in the actual use environment of the radiator are first recorded as three independent time series. The maximum and minimum normalization method is used to map the time series data of each environmental parameter to a unified range of 0 to 1 to generate a standardized environmental parameter vector. This vector consists of three components at any time point, representing the standardized salt spray concentration, temperature, and humidity at that moment. Subsequently, the Pearson correlation coefficient is calculated to analyze the linear correlation strength between each component time series of the standardized environmental parameter vector and the time series of the two corrosion characteristic channels (i.e., surface impedance characteristics and thermal resistance change rate characteristics) in the dual-channel corrosion characteristic dataset. The Pearson correlation coefficient quantifies the degree of linear correlation by calculating the covariance of the two sets of data divided by the product of their respective standard deviations. It ranges from -1 to 1 and is often used in statistics to represent the correlation between two variables. For any environmental parameter among salt spray concentration, temperature, and humidity, its correlation coefficient with the two corrosion characteristic channels is calculated respectively, and the average of their absolute values ​​is taken as the sensitivity coefficient of the environmental parameter. For any environmental parameter such as salt spray concentration , for calculating the sensitivity coefficient of the environmental parameter ,have: , in, , represents the Pearson correlation coefficient between the salt spray concentration time series in the standardized environmental parameter vector and the surface impedance characteristics in the dual-channel corrosion feature dataset; The Pearson correlation coefficient between the salt spray concentration time series and the thermal resistance change rate characteristic in the dual-channel corrosion characteristic dataset is shown in Figure 2. Similarly, the sensitivity coefficients of temperature and humidity in the environmental parameters are and Finally, based on the sensitivity coefficients of the determined environmental parameters, the final environmental sensitivity weight vector is constructed. Its components represent the relative importance of different environmental factors in inducing corrosion. By normalizing the three sensitivity coefficients so that their sum is 1, the weight of each environmental parameter can be obtained. For the calculation of the weight of salt spray concentration ,have: , Similarly, the temperature weight can be obtained and humidity weight These three weights together form the environmental sensitivity weight vector, enabling quantitative decoupling of the impact of complex environmental factors and identification of dominant factors. This allows for a scientific, data-driven determination of the contribution of different environmental stresses to the corrosion process of a specific radiator.

[0033] The generation of laboratory accelerated corrosion test spectrum and the acquisition of actual environmental fluctuation spectrum by long-term monitoring of the radiator include: Setting gradient-varying temperature and humidity data values ​​in the controllable salt spray environment to simulate corrosion processes under different environmental stress combinations and generate a laboratory accelerated corrosion test spectrum with corrosion rate variation trends; Analyze the recorded fluctuations of the salt spray concentration data value, the temperature data value, and the humidity data value to generate an actual environmental fluctuation spectrum; The laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum are normalized by time scale.

[0034] Specifically, generating a laboratory accelerated corrosion test spectrum requires a highly controlled laboratory environment. The temperature and humidity data within the test chamber are controlled by gradients to simulate diverse environmental stress combinations. Cyclic tests can be designed, encompassing multiple stages, such as high temperature and high humidity, high temperature and low humidity, and normal temperature and high humidity. Neutral salt spray of a specified concentration is continuously applied during each stage. During this period, dual-channel corrosion signature data is collected from the test heat sink: surface impedance time series data and thermal resistance change time series data. This generates not only corrosion data under a single, constant condition, but a time series that records the dynamic changes in corrosion rate in response to different environmental stresses, known as the laboratory accelerated corrosion test spectrum. Combined with the aforementioned extensive time series data accumulated through long-term monitoring of salt spray concentration, temperature, and humidity data around the heat sink in actual use, a spectrum of actual environmental fluctuations is constructed, reflecting the dynamic changes in the severity of the actual use environment. Finally, these two spectra, originating from distinct time scales, are time-normalized. Normalization aims to map them to a unified relative time coordinate system. For the calculation of the dimensionless time after normalization, ,have: , in, It is the absolute time from the start of the test or monitoring, which is directly provided by the timestamp of the data acquisition system. is the total duration of the laboratory accelerated corrosion test or actual environmental monitoring. This processing is applied to the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum respectively, so that the two time series are comparable on the normalized time axis despite their different total durations, laying the foundation for subsequent dynamic time warping alignment. Figure 2As shown in the figure, the horizontal axis represents the unification of the time scales of the laboratory and actual environment to the range of 0-1, which is convenient for direct comparison; the vertical axis represents the normalized corrosion rate, which reflects the relative change trend of corrosion intensity under different environments; the solid line is the laboratory accelerated corrosion test spectrum, which is a short-term, high-intensity corrosion environment simulated by the salt spray test chamber; the dotted line is the actual environment fluctuation spectrum, which represents the long-term monitoring of the temperature and humidity data of natural fluctuations in the real environment. By matching similar segments of the two curves, the corrosion equivalence relationship between the laboratory and the actual environment can be established, which is a key step in achieving accurate extrapolation from laboratory data to actual service life.

[0035] Optionally, generating time-domain aligned accelerated corrosion data includes: Performing time series segmentation on the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum to generate multiple local time windows; Under the constraint of the environmental sensitivity weight vector, calculating a similarity measure between the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum in each of the local time windows; Dynamically adjusting the scaling ratio of the time series according to the similarity metric value to keep the corrosion rate variation trend consistent with the actual environmental fluctuation spectrum; The laboratory accelerated corrosion test spectrum is interpolated and reconstructed based on the adjusted time series to generate time-domain aligned accelerated corrosion data.

[0036] Specifically, in order to generate the time-domain aligned accelerated corrosion data, the method adopts a weighted matching strategy based on a dynamic time warping algorithm. First, the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum that have been normalized by time scale are regarded as two time series to be matched. The time series segmentation of these two time series is actually to regard them as a sequence composed of discrete time points, where each time point is a local time window. Subsequently, under the constraint of the environmental sensitivity weight vector, the similarity measure between any two local time windows in the two time series is calculated. Construct a cost matrix whose matrix elements represent the cost or distance of matching the environmental state of a certain time point of the laboratory accelerated corrosion test spectrum with the environmental state of a certain time point of the actual environmental fluctuation spectrum. The calculation of this cost does not use the standard Euclidean distance, but introduces the environmental sensitivity weight. For the calculation of the representation of the laboratory accelerated corrosion test spectrum, the cost or distance of the actual environmental fluctuation spectrum is not the cost or distance of the actual environmental fluctuation spectrum. For the calculation of the environmental sensitivity weight vector, the cost of the laboratory accelerated corrosion test spectrum is not the cost or distance of the actual environmental fluctuation spectrum. The time point and the actual environmental fluctuation spectrum The weighted distance between time points have: , in, The spectrum of laboratory accelerated corrosion test is in Temperature data values, humidity data values, and salt spray concentration data values ​​recorded at each time point; The actual environmental fluctuation spectrum is Temperature data values, humidity data values, and salt spray concentration data values ​​recorded at each time point; is the weight value calculated based on the Pearson correlation coefficient. Then based on the weighted distance A cumulative cost matrix is ​​constructed, and an optimal path from the matrix's starting point to its end point is sought, minimizing the cumulative cost. This optimal path essentially defines a nonlinear mapping between the laboratory spectrum time axis and the actual environmental spectrum time axis. Horizontal movement in the path compresses the laboratory time axis to match a shorter actual environmental time axis, while vertical movement stretches the laboratory time axis to match a longer actual environmental time axis, thereby achieving dynamic scaling of the time series. The fundamental goal is to align the corrosion rate trends caused by a specific environmental combination in the laboratory with the fluctuation spectrum of a similar environmental combination in the real environment. Finally, reconstruction is performed based on the adjusted time series. Based on the optimal path obtained by dynamic time warping, the surface impedance and thermal resistance time series data from the laboratory accelerated corrosion test spectrum can be rearranged onto the time axis of the actual environmental spectrum according to the mapping relationship indicated by the path. Because the mapping is nonlinear, the rearranged data points may be uneven on the new time axis. Therefore, linear interpolation or spline interpolation methods are used for reconstruction, ultimately generating a time-continuous accelerated corrosion data series that is fully aligned with the actual environmental fluctuation spectrum in the time domain. This approach scientifically and nonlinearly maps the short, intense corrosion process in the laboratory to the long, fluctuating stress history of the real environment, generating a "virtual" corrosion evolution process that is synchronized with the real environment in time. This greatly improves the accuracy of the time scale conversion, thereby enhancing the reliability and accuracy of the entire life prediction framework.

[0037] Optionally, generating a basic life prediction value includes: Calculating the corrosion depth of the time-domain aligned accelerated corrosion data to generate a cumulative corrosion amount curve; Determining an initial life node according to an intersection of a preset corrosion threshold and the cumulative corrosion curve; The initial life node is linearly compensated in combination with the temperature data value and the humidity data value to generate a basic life prediction value.

[0038] Specifically, we firstly use the surface impedance time series data from the time-domain aligned accelerated corrosion data, especially the key parameter, charge transfer resistance, extracted from the electrochemical impedance spectroscopy. The corrosion rate is usually proportional to the corrosion current density, which in turn is related to the charge transfer resistance. Instantaneous corrosion rate at time ,have: , in, After time domain alignment The charge transfer resistance value at time t is directly obtained from the aligned surface impedance time series data. This constant is a comprehensive factor that incorporates the material's molar mass, density, the number of electrons transferred during the reaction, and the Faraday constant. This constant is determined by preparing standard specimens made of the same material as the radiator. These specimens are then placed in a salt spray chamber to simulate actual operating conditions, with a temperature and humidity gradient (e.g., 25°C / 80% RH, 35°C / 95% RH, etc.) and a fixed salt spray concentration (e.g., 5% NaCl). Electrochemical impedance spectroscopy (EIS) is measured using a three-electrode system, and an equivalent circuit is fitted to extract the charge transfer resistance. The actual corrosion rate is calculated using the acid wash weight loss method. The initial value is determined through linear regression, and the weighted average of multiple experimental sets is used to determine the value. Integrating this instantaneous corrosion rate along the time axis yields a cumulative corrosion curve, representing the total corrosion depth at any given moment. Compare the cumulative corrosion curve with the preset corrosion threshold. The first time point at which the total corrosion depth is not less than the corrosion threshold is the initial life node. The corrosion threshold is determined by laboratory accelerated corrosion tests to measure the corrosion depth or impedance change value when the radiator fails functionally (such as perforation, sudden increase in thermal resistance). The average value of multiple tests is taken as the threshold. For example, metallographic section analysis determines that the critical corrosion depth of fin perforation is 0.8mm. This node represents the time required for the radiator to reach its failure standard under the ideal model. Then, the systematic deviation introduced by the difference between the actual environmental average working conditions and the model reference working conditions is compensated and corrected. The compensation for the initial life node is based on the macroscopic deviation between the environmental average working conditions and the reference working conditions, which can be approximated as a linear relationship. The compensation combines the average value of the temperature data value and the humidity data value in the actual environmental fluctuation spectrum to generate a basic life prediction value. For calculating the basic life prediction value ,have: , in, is the initial lifespan node determined in the previous step; and They are calculated based on the actual environmental fluctuation spectrum, The average temperature value obtained from the temperature data and the average humidity value obtained from the humidity data within the time period; and It is the reference temperature and humidity data values ​​used when calibrating the model or defining the failure threshold; and is the temperature and humidity linear compensation coefficient. By comparing the laboratory accelerated test with the actual environmental monitoring data, the quantitative relationship between life and temperature and humidity deviation is established based on multiple linear regression to calculate the impact ratio of the average temperature and humidity deviation on life. Figure 3 As shown, the horizontal axis represents the total time history of the radiator from service start to failure; the vertical axis represents the total amount of corrosion damage accumulated on the radiator surface over time; the rising curve represents the cumulative corrosion depth over time; the horizontal dashed line represents the preset corrosion failure threshold, and the intersection is the initial life node (approximately 4.5 years); the gray area in the figure represents the failure risk zone after exceeding the threshold. By converting time-domain aligned dynamic corrosion data into a cumulative physical damage quantity, namely corrosion depth, a solid physical foundation is provided for life prediction. It is not just a data fit, but a simulation of the physical failure process. By setting a clear engineering failure threshold to determine the initial life node, the prediction results are directly linked to the actual engineering application requirements.

[0039] Optionally, the generation of the environment-structure coupling factor includes: Perform feature extraction based on the three-dimensional structural feature parameters to obtain surface curvature distribution data and weld density distribution data; performing a sensitivity analysis on the surface curvature distribution data according to the temperature data value and the humidity data value to generate a temperature-curvature correction coefficient and a humidity-curvature correction coefficient; Calculate and generate a thermal stress concentration factor based on the interaction between the weld density distribution data and the temperature data value; The temperature-curvature correction coefficient, the humidity-curvature correction coefficient and the thermal stress concentration factor are combined to construct an environment-structure coupling factor.

[0040] Specifically, firstly, a 3D laser scanner or industrial CT scanning equipment is used to perform non-contact measurement on the radiator to obtain its high-precision point cloud data; then, reverse engineering software (such as Geomagic or Imageware ) Convert the point cloud data into a three-dimensional CAD model to obtain the three-dimensional structural feature parameters; then mesh the three-dimensional structural feature parameters and extract features, calculate the main curvature of each surface grid unit, and generate a surface curvature distribution data covering the entire radiator surface. At the same time, through model recognition or manual annotation, determine the position, length and distribution of all welds, and quantify them as weld density distribution data, which characterizes the density of welds per unit area. For the surface curvature distribution data, its corrosion acceleration effect under different temperature and humidity conditions is studied. The residence time of condensate on surfaces with different curvatures and the degree of electrolyte enrichment under different temperatures and humidities are analyzed to fit the correction function of temperature and humidity on curvature accelerated corrosion, and generate temperature-curvature correction coefficient and humidity-curvature correction coefficient. For the calculation of temperature-curvature correction coefficient and humidity-curvature correction factor ,have: , in, Basic amplification factor, using flat specimens without curvature, corrosion tests are carried out at different temperatures and humidities to independently calibrate the basic amplification factor determined only by material properties; Temperature and humidity values ​​for environmental detection; For reference temperature and humidity, the temperature of 25°C and the relative humidity of 75%RH are used as the benchmark values; The surface curvature is the principal curvature data extracted through the three-dimensional structural characteristic parameters; To obtain the temperature curvature sensitivity parameter, a group of specimens with different curvatures were prepared and placed in a salt spray environment with gradient temperatures (25°C, 30°C, 35°C) at a fixed humidity (e.g. 75% RH). The corrosion rate of each specimen was measured. Fitting function , determined by the least squares method is the humidity curvature sensitivity parameter. Similarly, the temperature is fixed (such as 25°C) and it is fitted under gradient humidity. Since the material properties of the weld and its heat-affected zone are different from those of the radiator base material, local thermal stress will be generated in the weld area under temperature fluctuations. The temperature data values ​​in the actual environmental fluctuation spectrum are used as the thermal load input. The stress distribution of the area under different weld density distribution data is calculated through finite element analysis. The maximum stress value of the weld area and the average stress of the base material are extracted. The ratio of the two is used to obtain a thermal stress concentration factor that quantifies the stress concentration effect. Finally, the temperature-curvature correction coefficient, the humidity-curvature correction coefficient and the thermal stress concentration factor are combined to construct the environment-structure coupling factor. For constructing the environment-structure coupling factor ,have: , in, It is calculated based on the radiator's overall weld density distribution data and thermal stress concentration analysis, representing the impact of thermal stress on structural weaknesses. Since these three components are dimensionless risk amplification factors, their product constitutes a comprehensive quantification of corrosion risk. This factor is spatially distributed, with high-value areas corresponding to potential failure hotspots in the radiator. It can provide targeted guidance on basic life predictions, elevating the prediction from an "average" lifespan to an accurate assessment of the "individual" lifespan of specific structures. This makes the final prediction more realistic and effectively captures the risk of early failure introduced by structural design.

[0041] Optionally, generating a final lifespan prediction result includes: Perform correlation mapping based on the environment-structure coupling factor to generate a life compensation coefficient; Performing nonlinear correction on the basic life prediction value according to the life compensation coefficient to generate a corrected life prediction value; Smoothing is performed based on the corrected life prediction value to generate a final life prediction result.

[0042] Specifically, the final life prediction result is generated on top of the basic life prediction value, and is a final refinement process that further incorporates the structural specificity effect. First, the environment-structure coupling factor is converted into a correction multiplier that can directly act on the life prediction value, that is, the life compensation coefficient. This mapping relationship must reflect physical reality, that is, the stronger the environment-structure coupling effect, the more significant the actual life is shortened. The correlation mapping is characterized by an exponential decay model, which can effectively describe the characteristics of corrosion-accelerated failure. For calculating the life compensation coefficient ,have: , in, It is a positive mapping sensitivity parameter, and its value is calibrated by performing accelerated corrosion tests on radiators with different structural characteristics and fitting regression analysis with their actual failure life data, using maximum likelihood estimation or Bayesian optimization. Then, the basic life prediction value is nonlinearly corrected according to the life compensation coefficient. The life compensation coefficient is used as a weight to adjust the basic life prediction value to obtain a revised life prediction value that takes into account the structural influence. For calculating the revised life prediction value ,have: , in, It is the basic life prediction value calculated in the previous step, which mainly reflects the dynamic cumulative effect of environmental factors. , The life span will be effectively shortened, and the degree of shortening is determined by the structural weaknesses of the radiator itself and the intensity of its interaction with the environment. The above calculation process may introduce high-frequency fluctuations caused by data noise or model jumps. In order to obtain a stable prediction curve with more engineering application value, smoothing is required. Algorithms such as moving average filtering or Savitzky-Golay filtering can be used. If the moving average method is used, the final prediction result is the average value of the corrected life prediction value within a time window. For calculation The final smoothed prediction result at time ,have: , in, for The final life prediction result output at every moment, The size of the smoothing window should be a balance between filtering out noise and retaining the true lifespan variation trend; Express arrive This step transforms a series of discrete or fluctuating prediction points into a smooth life prediction curve. This can deeply reflect life variations caused by design differences such as radiator fin density, bend curvature, and weld distribution. The prediction result is no longer an average value that ignores individual structural characteristics, but rather a highly customized dynamic assessment that is tightly coupled to the specific product structure.

[0043] Optionally, the method further includes: Obtaining a state record of the radiator when it actually fails, and combining the surface impedance time series data and the thermal resistance change time series data to generate actual failure data; Comparing the final life prediction result with the actual failure data to generate a prediction error distribution graph; The environmental sensitivity weight vector is reversely adjusted according to the prediction error distribution map, and the calculation rule of the environment-structure coupling factor is updated.

[0044] Specifically, when a heat sink deployed in the field eventually fails due to corrosion, the precise time of failure must be recorded in detail. Failure analysis methods, such as metallographic section analysis or scanning electron microscopy, can be used to determine the specific failure state, including corrosion depth, perforation, or performance degradation in key areas. This data is combined with surface impedance time series data and thermal resistance change time series data to construct a complete set of actual failure data. The predicted failure time in the final life prediction results is compared with the recorded actual failure time, and the prediction error between the two is calculated. This process is repeated for a batch of failed heat sink samples to generate a series of prediction error values. These error values ​​are statistically analyzed to generate a prediction error distribution graph. This graph, typically in the form of a histogram, can intuitively demonstrate the overall performance of the prediction model, including whether there is systematic overestimation or underestimation, as well as the degree of dispersion in the prediction results. Finally, based on the information revealed by the prediction error distribution graph, the prediction model is adjusted and optimized. If the mean of the error distribution graph deviates significantly from zero, it indicates that the model has systematic bias, and the environmental sensitivity weight vector needs to be adjusted. First, based on the prediction error distribution map, gradient descent or Bayesian optimization algorithm is used to iteratively adjust the parameters in the environmental sensitivity weight vector (such as increasing the humidity weight by 3%) with the goal of minimizing the average prediction error. Second, for the update of the environment-structure coupling factor, the sensitivity parameters in the temperature-curvature correction coefficient and the humidity-curvature correction coefficient are refitted through regression analysis (such as adjusting the sensitivity parameters in the exponential decay model). The FEA model's boundary conditions (such as the load spectrum for weld thermal stress analysis) are then modified based on the new failure data to ensure consistency between model parameters and the physical failure mechanism. For example, if the model consistently underestimates lifetime in high-humidity environments, the humidity parameter's weight in the weight vector should be increased. This adjustment process can be accomplished by introducing an optimization algorithm to iteratively update the weight parameters with the goal of minimizing the average error across all cases. Furthermore, if a strong correlation between the prediction error and specific structural characteristics of the heat sink (such as failure consistently occurring at welds) is observed, the calculation rules for the environmental-structural coupling factor need to be updated. This involves revising the calculation formulas for its component factors (such as the temperature-curvature correction factor and the thermal stress concentration factor) or adjusting the nonlinear weighting exponents used to combine these factors to more accurately quantify the life-degrading effects of specific structural vulnerabilities. This data-driven self-calibration process ensures that the accuracy of the prediction model continues to improve with the amount of failure data collected, ultimately achieving long-term, high-precision predictions of the salt spray corrosion lifetime of heat sinks.

[0045] Based on the same inventive concept, Figure 4 As shown, the present invention also provides a radiator salt spray corrosion life prediction system based on dynamic time warping, the system comprising: The dual-channel data module is used to obtain the surface impedance time series data and thermal resistance change time series data of the radiator in a salt spray environment to generate a dual-channel corrosion feature data set; The environmental detection module is used to deploy a multi-sensor network in the actual use environment of the radiator to record salt spray concentration data values, temperature data values ​​and humidity data values; A weight vector construction module, configured to establish an environmental sensitivity weight vector based on the dual-channel corrosion feature data set, the salt spray concentration data value, the temperature data value, and the humidity data value; a corrosion simulation generation module, configured to simulate accelerated corrosion in the salt spray environment based on the temperature data value and the humidity data value, generate a laboratory accelerated corrosion test spectrum, and collect an actual environmental fluctuation spectrum by long-term monitoring of the radiator; A time domain alignment processing module, configured to match and align the laboratory accelerated corrosion test spectrum with the time periods having equivalent corrosion effects in the actual environmental fluctuation spectrum based on the environmental sensitivity weight vector, to generate time domain aligned accelerated corrosion data; A basic life prediction module, configured to generate a basic life prediction value by performing calculations based on the time-domain aligned accelerated corrosion data; a structural coupling calculation module, configured to obtain three-dimensional structural characteristic parameters of the radiator, and calculate and generate an environment-structure coupling factor by combining the interaction between the three-dimensional structural characteristic parameters and the temperature data value and the humidity data value; The life dynamic correction module is used to correct the basic life prediction value through the environment-structure coupling factor to generate a final life prediction result.

[0046] To verify the feasibility of this invention, it was applied to a large-scale chemical enterprise in a coastal region. This enterprise's critical production process relies on a large-scale heat exchange system. Its core component, the radiator, is exposed to a corrosive environment of high salt spray and high humidity for a long time, resulting in frequent failures and impacting production safety and efficiency. The enterprise hoped to use the method of this invention to dynamically predict the corrosion life of its critical radiators in service, thereby enabling predictive maintenance and avoiding unplanned downtime.

[0047] In this example, the chemical company selected an HX-500 radiator from its cooling tower system as the monitoring target. First, using the dual-channel data module, an electrochemical three-electrode system and a high-precision temperature sensor array were deployed in key corrosion areas of the radiator. This system automatically collected electrochemical impedance spectroscopy and thermal resistance data every six hours, generating a dual-channel corrosion signature dataset. Simultaneously, the environmental monitoring module, using a multi-sensor network deployed around the radiator, recorded real-time time series data on salt spray concentration, ambient temperature, and humidity.

[0048] During the initial implementation phase, the system collected three months of field data. Based on this data, the weight vector construction module normalized the environmental data and, through Pearson correlation analysis, calculated the sensitivity coefficients between each environmental parameter and the dual-channel corrosion signature data. For example, the analysis found that humidity had the strongest correlation with corrosion rate, with a sensitivity coefficient of 0.85, while temperature had a sensitivity coefficient of 0.62 and salt spray concentration had a sensitivity coefficient of 0.45. Based on this calculation, an environmental sensitivity weight vector was derived, with a weight of 0.44 for humidity, 0.32 for temperature, and 0.24 for salt spray concentration.

[0049] In parallel, the corrosion simulation generation module conducted accelerated corrosion testing on a new HX-500 radiator in a controlled environment chamber. By setting a gradient temperature and humidity combination, it generated a laboratory accelerated corrosion test spectrum. Subsequently, under the constraints of the environmental sensitivity weight vector, the time domain alignment processing module used a dynamic time warping algorithm to perform nonlinear time domain alignment between the 300-hour laboratory accelerated spectrum and the three-month actual environmental fluctuation spectrum. This generated an hourly accelerated corrosion rate sequence synchronized with the actual environmental timeline.

[0050] The basic life prediction module integrates this time-aligned corrosion rate series to generate a cumulative corrosion curve. Based on the failure criterion for this heat sink model (a 15% increase in total thermal resistance), a corrosion threshold is determined. The intersection of the curve and the threshold gives an initial lifespan of 24.5 months. Combined with the average temperature and humidity during this period, a linear compensation formula is used to calculate the basic lifespan prediction value to be 22.8 months.

[0051] The Structural Coupling Calculation Module then imported a 3D design model of the HX-500 radiator and identified the highly curved edges of the fins and the welds at the tube-to-sheet connections as structural weaknesses. Finite element simulation, combined with actual temperature fluctuation data, determined the maximum environmental-structural coupling factor at these weak points to be 1.42. The Dynamic Life Correction Module mapped this factor to a life compensation coefficient of 0.704 and, applying nonlinear correction, generated a final life prediction of 17.5 months. This result was significantly lower than the baseline life prediction, highlighting the decisive influence of structural weaknesses on life.

[0052] After 18 months of system operation, a monitored HX-500 radiator failed due to a leak. Its actual lifespan of 18 months closely matched the 17.5-month prediction based on the present invention, with a relative error of only 2.8%. This actual failure data was fed into the system, and through a feedback correction mechanism, fine-tuned the calculation rules for the environmental sensitivity weight vector and the environment-structure coupling factor, further improving the accuracy of subsequent predictions.

[0053] Table 1 Radiator environment and corrosion characteristics monitoring data

[0054] Table 2 Radiator life prediction calculation data table

[0055] Table 3 Comparison of life prediction results and actual failure and model optimization data

[0056] Tables 1-3 above record the actual application data of the present invention in the chemical enterprise, and show in detail the performance of the system in data acquisition, quantitative calculation and prediction verification.

[0057] As shown in Table 1, the system successfully collected multi-dimensional real-time data on the environment and corrosion status, clearly showing the trend of decreasing charge transfer resistance and increasing rate of change of thermal resistance, which represent the degree of corrosion, under high humidity and high salt fog conditions, providing reliable data input for subsequent analysis.

[0058] Table 2 illustrates the core calculation process of our invention. The significant difference between the baseline life prediction and the final life prediction (for example, for HX-500-01, the correction from 22.8 months to 17.5 months) intuitively demonstrates the necessity and effectiveness of introducing the environment-structure coupling factor. It successfully quantifies the accelerated corrosion effect of structural weaknesses under specific environments, bringing the prediction closer to physical reality.

[0059] Table 3 demonstrates the high accuracy of the proposed method by comparing the predicted results with actual failure data. The -2.8% prediction error significantly outperforms traditional empirical estimation methods. More importantly, the system leverages this real-world data point to complete closed-loop feedback and adaptive optimization of the model, demonstrating the proposed method's ability to self-learn and continuously evolve, ensuring long-term reliability.

[0060] It should be noted that the electrical connections between the above-mentioned units do not necessarily mean direct connections of lines. Indirect connections are applicable to the embodiments of the present invention as long as the purpose of the present invention is achieved. The above description is only an exemplary embodiment of the present invention and is not intended to limit the scope of the present invention.

[0061] That is, any equivalent changes and modifications made according to the teachings of the present invention are still within the scope of the present invention. Those skilled in the art will readily conceive of other embodiments of the present invention after considering the disclosure of the specification and practical truths. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary technical means in the art not described herein.

Claims

1. A radiator salt spray corrosion life prediction method based on dynamic time warping is characterized by: The method comprises: Obtain the surface impedance time series data and thermal resistance change time series data of the radiator in a salt spray environment to generate a dual-channel corrosion feature data set; Deploy a multi-sensor network in the actual use environment of the radiator to record salt spray concentration data values, temperature data values, and humidity data values; Establishing an environmental sensitivity weight vector based on the dual-channel corrosion feature data set, the salt spray concentration data value, the temperature data value, and the humidity data value; Simulating accelerated corrosion in the salt spray environment based on the temperature data value and the humidity data value to generate a laboratory accelerated corrosion test spectrum, and collecting an actual environmental fluctuation spectrum by long-term monitoring of the radiator; Matching and aligning the laboratory accelerated corrosion test spectrum with time periods having equivalent corrosion effects in the actual environmental fluctuation spectrum based on the environmental sensitivity weight vector to generate time-domain aligned accelerated corrosion data; Perform calculations based on the time-domain aligned accelerated corrosion data to generate a basic life prediction value; Acquire three-dimensional structural characteristic parameters of the radiator, and calculate and generate an environment-structure coupling factor by combining the interaction between the three-dimensional structural characteristic parameters and the temperature data value and the humidity data value; The basic life prediction value is corrected by the environment-structure coupling factor to generate a final life prediction result.

2. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 1 is characterized in that: Generating a dual-channel corrosion feature data set includes: Collect continuous measurement values ​​of the electrochemical impedance spectrum of the radiator surface over time to generate a surface impedance time series graph; Simultaneously monitor the heat conduction performance decay rate of the radiator and generate a thermal resistance change rate sequence; The surface impedance time series graph and the thermal resistance change rate sequence are subjected to time stamp synchronization fusion processing to generate a dual-channel corrosion feature data set.

3. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 1 is characterized in that: The establishing of the environmental sensitivity weight vector includes: Normalizing the salt spray concentration data value, the temperature data value, and the humidity data value to generate a standardized environmental parameter vector; Analyzing the correlation between the standardized environmental parameter vector and the dual-channel corrosion characteristic data set through the Pearson correlation coefficient to determine the sensitivity coefficient of each environmental parameter; The weight relationship among the salt fog concentration data value, the temperature data value, and the humidity data value is integrated based on the sensitivity coefficient and quantized to establish an environmental sensitivity weight vector.

4. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 1 is characterized in that: The generation of laboratory accelerated corrosion test spectrum and the acquisition of actual environmental fluctuation spectrum by long-term monitoring of the radiator include: Setting gradient-varying temperature and humidity data values ​​in the controllable salt spray environment to simulate corrosion processes under different environmental stress combinations and generate a laboratory accelerated corrosion test spectrum with corrosion rate variation trends; Analyze the recorded fluctuations of the salt spray concentration data value, the temperature data value, and the humidity data value to generate an actual environmental fluctuation spectrum; The laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum are normalized by time scale.

5. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 4 is characterized in that: The generating of time-domain aligned accelerated corrosion data comprises: Performing time series segmentation on the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum to generate multiple local time windows; Under the constraint of the environmental sensitivity weight vector, calculating a similarity measure between the laboratory accelerated corrosion test spectrum and the actual environmental fluctuation spectrum in each of the local time windows; Dynamically adjusting the scaling ratio of the time series according to the similarity metric value to keep the corrosion rate variation trend consistent with the actual environmental fluctuation spectrum; The laboratory accelerated corrosion test spectrum is interpolated and reconstructed based on the adjusted time series to generate time-domain aligned accelerated corrosion data.

6. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 5 is characterized in that: Generating a basic life prediction value includes: Calculating the corrosion depth of the time-domain aligned accelerated corrosion data to generate a cumulative corrosion amount curve; Determining an initial life node according to an intersection of a preset corrosion threshold and the cumulative corrosion curve; The initial life node is linearly compensated in combination with the temperature data value and the humidity data value to generate a basic life prediction value.

7. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 1 is characterized in that: The generation environment-structure coupling factors include: Perform feature extraction based on the three-dimensional structural feature parameters to obtain surface curvature distribution data and weld density distribution data; performing a sensitivity analysis on the surface curvature distribution data according to the temperature data value and the humidity data value to generate a temperature-curvature correction coefficient and a humidity-curvature correction coefficient; Calculate and generate a thermal stress concentration factor based on the interaction between the weld density distribution data and the temperature data value; The temperature-curvature correction coefficient, the humidity-curvature correction coefficient and the thermal stress concentration factor are combined to construct an environment-structure coupling factor.

8. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 7 is characterized in that: Generating the final life prediction result includes: Perform correlation mapping based on the environment-structure coupling factor to generate a life compensation coefficient; Performing nonlinear correction on the basic life prediction value according to the life compensation coefficient to generate a corrected life prediction value; Smoothing is performed based on the corrected life prediction value to generate a final life prediction result.

9. The method for predicting the salt spray corrosion life of a radiator based on dynamic time warping according to claim 8 is characterized in that: The method further comprises: Obtaining a state record of the radiator when it actually fails, and combining the surface impedance time series data and the thermal resistance change time series data to generate actual failure data; Comparing the final life prediction result with the actual failure data to generate a prediction error distribution graph; The environmental sensitivity weight vector is reversely adjusted according to the prediction error distribution map, and the calculation rule of the environment-structure coupling factor is updated.

10. A radiator salt spray corrosion life prediction system based on dynamic time warping, applied to a radiator salt spray corrosion life prediction method based on dynamic time warping as claimed in any one of claims 1 to 9, characterized in that: The system comprises: The dual-channel data module is used to obtain the surface impedance time series data and thermal resistance change time series data of the radiator in a salt spray environment to generate a dual-channel corrosion feature data set; The environmental detection module is used to deploy a multi-sensor network in the actual use environment of the radiator to record salt spray concentration data values, temperature data values ​​and humidity data values; A weight vector construction module, configured to establish an environmental sensitivity weight vector based on the dual-channel corrosion feature data set, the salt spray concentration data value, the temperature data value, and the humidity data value; a corrosion simulation generation module, configured to simulate accelerated corrosion in the salt spray environment based on the temperature data value and the humidity data value, generate a laboratory accelerated corrosion test spectrum, and collect an actual environmental fluctuation spectrum by long-term monitoring of the radiator; A time domain alignment processing module, configured to match and align the laboratory accelerated corrosion test spectrum with the time periods having equivalent corrosion effects in the actual environmental fluctuation spectrum based on the environmental sensitivity weight vector, to generate time domain aligned accelerated corrosion data; A basic life prediction module, configured to generate a basic life prediction value by performing calculations based on the time-domain aligned accelerated corrosion data; a structural coupling calculation module, configured to obtain three-dimensional structural characteristic parameters of the radiator, and calculate and generate an environment-structure coupling factor by combining the interaction between the three-dimensional structural characteristic parameters and the temperature data value and the humidity data value; The life dynamic correction module is used to correct the basic life prediction value through the environment-structure coupling factor to generate a final life prediction result.

Images