Comparison and analysis optimization method based on cubic spline interpolation dynamic and static detection data

Abstract

The application provides a comparative analysis optimization method based on cubic spline interpolation dynamic and static detection data, and relates to the technical field of electric digital data processing. The fatigue characteristic parameter set extracted is smoothed by the cubic spline interpolation method, continuous interpolation curve data is generated, the continuous fatigue characteristic interpolation data is synchronously fused with the dynamic and static detection data, a multi-dimensional comprehensive data set is constructed, and a more abundant and reliable data basis is provided for a fault prediction model. Furthermore, the fault prediction model based on the continuous interpolation data and the comprehensive detection data can more effectively identify early signs and potential risks of material fatigue, provide scientific maintenance and replacement strategies, and significantly improve the safety and reliability of engineering structures.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing technology, specifically to a comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data. Background Technology

[0002] In existing technologies, material fatigue testing primarily relies on a combination of high-definition image acquisition and ultrasonic echo analysis. This approach assesses the fatigue state of the material by real-time monitoring of surface and internal defects on the test piece. For example, ultrasonic testing can penetrate deep into the material, revealing invisible micro-cracks or voids, while high-definition image acquisition provides intuitive information about surface defects. In recent years, with the development of computer image processing and data analysis algorithms, numerical methods such as cubic spline interpolation have been widely applied to smoothing and fitting fatigue data to improve the continuity and accuracy of the test data.

[0003] In the prior art, publication number CN119720871A, entitled "A Data Optimization Method, Apparatus, Equipment, and Medium Based on Cubic Spline Interpolation," the method includes acquiring the original pressure data of an airfoil model surface after a wind tunnel pressure test; generating a pressure data matrix based on the original pressure data; calculating the pressure coefficient of each pressure measurement point in the pressure data matrix to obtain a pressure coefficient matrix for all pressure measurement points; setting a corresponding interpolation angle sequence for each pressure measurement point; interpolating the determinant of the pressure coefficients of the pressure measurement points in the pressure coefficient matrix to the corresponding interpolation angle sequence using the cubic spline interpolation method to obtain the pressure coefficient interpolation results for all pressure measurement points of the airfoil model; and using the pressure coefficient interpolation results to optimize the wind tunnel pressure test data, which can avoid the Runge phenomenon caused by higher-order Lagrange interpolation polynomials, solve the data oscillation problem at the edge of the interpolation interval, improve the data rounding accuracy, and provide technical support for the optimization of wind tunnel pressure tests.

[0004] Existing technologies still have many shortcomings in the process of material fatigue detection and data analysis:

[0005] 1. First, traditional fatigue feature extraction methods often rely on discrete sampling point data, lacking a continuous representation of defect feature parameters, resulting in insufficient accuracy and reliability in subsequent fatigue state assessment and fault prediction. Second, existing data smoothing and interpolation methods often struggle to maintain the smoothness and fitting accuracy of interpolation curves when dealing with high-noise or incomplete data, affecting the effectiveness of subsequent comprehensive data analysis.

[0006] 2. In addition, the method for constructing a fault prediction model based on continuous fatigue feature interpolation data and dynamic and static test data is still immature, making it difficult to fully utilize the crack propagation risk level and stress concentration of the test piece for further adjustments in fault prediction and risk assessment.

[0007] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide a comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] The method for comparative analysis and optimization based on cubic spline interpolation dynamic and static detection data includes the following steps:

[0011] Step S1: Collect high-definition image data and ultrasonic echo data of multiple fixed sampling points on the surface of the test piece under a fixed environment, and analyze these data to generate material fatigue image data for each fixed sampling point; at the same time, collect dynamic and static test data of the test piece.

[0012] Step S2: Extract the defect feature information of the material fatigue image data from each fixed sampling point to form a fatigue feature parameter set;

[0013] Furthermore, by applying cubic spline interpolation, the fatigue characteristic parameter set is smoothed to form continuous interpolation curve data, thereby obtaining continuous fatigue characteristic interpolation data.

[0014] Step S3: Synchronously fuse the continuous fatigue feature interpolation data with the dynamic and static detection data to form a comprehensive dataset;

[0015] Step S4: Receive the comprehensive dataset and construct a fault prediction model based on continuous fatigue feature interpolation data and dynamic and static detection data;

[0016] Step S5: Obtain the fatigue characteristic parameter set and perform comprehensive weighted calculation and analysis to construct a defect propagation risk index. The defect propagation risk index is used to assess the crack propagation risk level of the test piece.

[0017] Step S6: Obtain the local stress values ​​and overall stress values ​​at multiple fixed sampling points on the test piece, determine the stress concentration of the test piece by calculating the ratio, and obtain the potential risk assessment results for each fixed sampling point;

[0018] Step S7: Based on the crack propagation risk level assessment results and potential risk assessment results of the test piece, provide correction strategies for the output results of the fault prediction model, and apply the optimized output results to the fault prediction comparison process of the test piece in the next testing period.

[0019] Compared with the prior art, the beneficial effects of the present invention are: by smoothing the extracted fatigue feature parameter set through cubic spline interpolation, continuous interpolation curve data is generated, overcoming the limitations caused by data discretization in traditional methods; at the same time, the continuous fatigue feature interpolation data is synchronously integrated with dynamic and static detection data to construct a multi-dimensional comprehensive dataset, providing a richer and more reliable data foundation for the fault prediction model;

[0020] Furthermore, the fault prediction model based on continuous interpolation data and comprehensive test data, by making full use of the crack propagation risk level and stress concentration of the test component to further adjust the fault prediction and risk assessment, can more effectively identify early signs and potential risks of material fatigue, provide scientific maintenance and replacement strategies, and significantly improve the safety and reliability of engineering structures. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall method flow of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0023] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0024] Example 1:

[0025] Please see Figure 1 The present invention provides a technical solution:

[0026] Example 1:

[0027] A comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data specifically includes:

[0028] Step S1: Collect high-definition image data and ultrasonic echo data of multiple fixed sampling points on the surface of the test piece under a fixed environment, and analyze these data to generate material fatigue image data for each fixed sampling point; at the same time, collect dynamic and static test data of the test piece.

[0029] Further explanation: The acquisition of continuous high-definition image data and ultrasonic echo data of the surface of the test piece under a fixed environment, along with the acquisition of dynamic and static test data of the test piece, specifically includes:

[0030] The component under test in this embodiment is an engine piston;

[0031] Under a fixed environment, the stress-bearing area on the surface of the workpiece is selected as the detection area; in this embodiment, the fixed environment is a temperature of 20℃±2℃ and a humidity of 50%±5%.

[0032] Using a scanning electron microscope of model "SEM-ModelX", images were acquired at N fixed sampling points distributed in the detection area to obtain high-resolution image data for each fixed sampling point; the index of the fixed sampling point was denoted as i, and i∈{1,2,…,N}; where i represents the index of the fixed sampling point; in this embodiment, N≥10; and the resolution of each sampling point image was ensured to be no less than 2048×2048 pixels;

[0033] Simultaneously, an ultrasonic testing instrument of model "UT-ModelY" was used to obtain ultrasonic echo data from at least three repeated measurements at each fixed sampling point;

[0034] Further explanation: The dynamic and static test data include the vibration index and temperature index of the test piece;

[0035] When the component under test is an engine piston:

[0036] Vibration indicators monitor the frequency and amplitude of engine piston vibration during operation; this type of data is suitable for detecting mechanical faults and wear.

[0037] Temperature parameters monitor changes in engine piston temperature. This type of data is used to identify overheating problems and potential thermal fatigue.

[0038] Data acquisition is conducted in both dynamic and static states; the dynamic state is during the operation of the engine piston equipment, and the static state is during the shutdown and maintenance of the engine piston equipment.

[0039] Vibration parameters include vibration frequency index. and vibration amplitude index ;

[0040] Define the vibration frequency index By measuring the vibration frequency of the test component during operation and normalizing it to reflect its state within a safe range; this vibration frequency index... The closer it is to 1, the higher the vibration frequency, which indicates a greater risk of mechanical failure or wear of the test piece.

[0041] The specific steps for calculating the vibration frequency index are as follows:

[0042] Obtain the vibration frequency f and the maximum threshold of the vibration frequency of the test piece. ;

[0043] Maximum threshold of vibration frequency Defined as the highest safe frequency of the device under test under normal operating conditions;

[0044] Vibration frequency index By comparing the actual vibration frequency f with the maximum vibration frequency threshold Perform ratio calculations and adjust using an exponential function to ensure that the output value range is within the interval (0,1);

[0045]

[0046] In this embodiment, when the component under test is an aero-engine piston, when When the value approaches 1, it indicates that the piston's vibration frequency is close to the design or safety limit, which can lead to mechanical failures such as bearing damage or gear wear. Therefore, this index helps to identify abnormal conditions in a timely manner and ensure the reliable operation of the engine piston.

[0047] Define vibration amplitude index The severity of vibration of the test piece is assessed by measuring and normalizing the amplitude of piston vibration; the vibration amplitude index. The higher the value, the greater the risk of mechanical failure or wear of the part under test;

[0048] Vibration Amplitude Index The specific calculation steps are as follows:

[0049] Obtain the vibration amplitude A and the maximum vibration amplitude threshold of the test piece. ;

[0050] Vibration Amplitude Index By comparing the actual vibration amplitude A with the maximum vibration amplitude threshold Perform ratio calculations to ensure that the output value range is between (0,1);

[0051]

[0052] In this embodiment, when the test component is an aero-engine piston, the vibration amplitude index is... Approaching or reaching the maximum vibration amplitude threshold This indicates that the structure is loose or the components are unbalanced;

[0053] By using a weighted average method, combined with the vibration frequency index and vibration amplitude index Generate a comprehensive vibration index ; The range is within the interval (0,1);

[0054] Define the comprehensive vibration index The calculation formula is as follows:

[0055]

[0056] in, It is the vibration frequency index Weighting coefficients; It is the vibration amplitude index The weighting coefficients; satisfying .

[0057] The closer the value is to 1, the higher the risk of failure of the device under test.

[0058] Further explanation: Setting temperature parameters includes temperature index. and temperature change rate index ;

[0059] Temperature Index By measuring and normalizing the temperature of the test piece, the impact of temperature conditions on thermal fatigue and overheating of the test piece material is assessed; temperature index. The higher the value, the higher the temperature, and the greater the potential risk of thermal fatigue and overheating;

[0060] Temperature Index The specific calculation steps are as follows:

[0061] Set maximum temperature threshold This represents the highest safe temperature under normal operating conditions of the device under test.

[0062] Temperature Index By comparing the actual temperature T with the maximum temperature threshold Perform ratio calculations to ensure that the output value range is within the interval (0,1);

[0063]

[0064] Based on the test component being an engine piston, when When the value approaches 1, it indicates that the temperature of the engine piston is close to or has reached the safety limit, which may lead to thermal fatigue or overheating of the material under test, thereby affecting the material performance and lifespan.

[0065] Temperature change rate index The impact of temperature change rate on thermal shock and structural damage of the test specimen is assessed by measuring and normalizing the rate of temperature change over a period of time; the temperature change rate index. The higher the value, the faster the temperature changes and the greater the potential risk of thermal shock.

[0066] Set the maximum threshold for temperature change rate The unit is degrees Celsius per second (°C / s), which is defined as the highest safe rate of change of the test piece under normal operating conditions. Indicates the test piece during a certain period of time Temperature change value;

[0067] Temperature change rate index By using the actual temperature change rate Maximum threshold of temperature change Perform ratio calculations to ensure that the output value range is within the interval (0,1).

[0068]

[0069] when When the value approaches 1, it indicates that the faster the temperature of the test piece changes, the more thermal shock it will cause, and the greater the damage to the material structure of the test piece.

[0070] By using a weighted average method, combined with the temperature index and temperature change rate index Generate a comprehensive temperature index The comprehensive temperature index The range is within the interval (0,1);

[0071] Comprehensive temperature index The calculation formula is as follows:

[0072]

[0073] in, It is a temperature index Weighting coefficients; It is the temperature change rate index The weighting coefficients; satisfying ;

[0074] Comprehensive temperature index The closer the value is to 1, the higher the degree of thermal fatigue of the test piece.

[0075] The logic behind the changes in vibration index parameters is as follows:

[0076] As the vibration frequency f increases: It increases linearly with increasing f;

[0077] An increase in vibration frequency indicates that the bearings of the engine piston corresponding to the test component are damaged or the gears are worn more severely, leading to an increased risk of mechanical failure.

[0078] When f increases The closer the value is to 1, the closer the vibration frequency is to or the maximum safety threshold is reached, indicating that the test piece needs to be checked immediately.

[0079] The vibration amplitude A increases: The amplitude increases linearly with increasing A; an increase in vibration amplitude indicates that the test piece is loose or the component is unbalanced, and further inspection is required to prevent serious failure.

[0080] When A increases The closer the value is to 1, the closer the vibration amplitude is to the maximum threshold, indicating a structural problem in the test piece.

[0081] Comprehensive Vibration Index Increase: Follow and / or The increase in the overall vibration index reflects the deterioration of the overall vibration characteristics, indicating the need for timely maintenance to prevent the fault from escalating.

[0082] The logic behind the changes in temperature parameters is as follows:

[0083] Temperature T increases: It increases linearly with increasing temperature (T); rising temperature can lead to thermal fatigue or overheating of the material, affecting its properties and lifespan; as T increases... The closer the value is to 1, the more it indicates that the temperature is approaching the maximum safe threshold and cooling measures need to be taken.

[0084] Temperature change rate index Increase: Follow Increase linearly; rapid temperature change rates can trigger thermal shock, damaging the material structure; when When increasing, The closer the value is to 1, the faster the temperature changes, which can lead to thermal fatigue and structural damage to the test piece.

[0085] Comprehensive temperature index Increase: With temperature index and temperature change rate index The increase in the comprehensive temperature index reflects the deterioration of the overall temperature condition, indicating the need to take timely measures to prevent material damage; when An increase indicates that both the temperature and the rate of temperature change are approaching or have reached the threshold, requiring immediate cooling and maintenance measures.

[0086] It should be noted that the "thresholds" of the above indices were determined by the expert group using the fuzzy hierarchical analysis method (FAHP), which will not be elaborated here. It is necessary to ensure that the "thresholds" of each index are set within the corresponding index value range.

[0087] Further explanation: Generating material fatigue image data for each fixed sampling point specifically includes:

[0088] 1.1) Preprocess the high-resolution image data of each fixed sampling point and the ultrasonic echo data from at least three repeated measurements. Based on each repeated measurement, generate a preliminary defect distribution map for each fixed sampling point; wherein:

[0089] 1.11) Preprocessing includes noise filtering, using wavelet denoising algorithms to remove high-frequency noise and ensure signal smoothness and continuity; the specific steps are as follows:

[0090] For each fixed sampling point, ultrasonic echo data were collected from at least three repeated measurements using an ultrasonic testing instrument.

[0091] Discrete Wavelet Transform (DWT) was used to decompose the ultrasonic echo data of each measurement.

[0092] The signal was decomposed using the Daubechies 4 (db4) wavelet basis, down to level 3.

[0093] For each level of detail coefficients, a soft threshold denoising method is applied, with the threshold set to 3 times the standard deviation of the noise.

[0094] Reconstruct the denoised signal to ensure its smoothness and continuity.

[0095] For each fixed sampling point, the three denoised signals are overlapped and averaged to obtain the averaged echo signal.

[0096] 1.12) The time reversal and concentrated echo imaging algorithm is applied to process the preprocessed ultrasonic echo data to obtain the depth information for imaging the internal cracks of the test piece.

[0097] The specific steps are as follows:

[0098] 1. Time reversal processing:

[0099] Perform a time reversal operation on the averaged echo signal to obtain the inverted signal at a fixed sampling point i. This involves reversing the signal sequence in chronological order.

[0100] 2. Concentrated echo imaging:

[0101] The inverted signal at fixed sampling point i Input-focused echo imaging algorithm (Matched Filter Technique).

[0102] The response function of the matched filter is set to a preset defect feature template, and the parameters of the matched filter are set to the material properties of the test piece and the expected crack size.

[0103] The signal is matched with the matched filter through convolution operation, and the defect echo component in the signal is extracted.

[0104] Reconstruct the temporal image to generate a basic defect distribution map at a fixed sampling point i. This reflects the depth information of the internal cracks in the test piece.

[0105] 1.13) Perform edge detection and feature extraction on the high-definition image data of each fixed sampling point i, and mark the cracks or defects qualitatively or quantitatively on the image to obtain two-dimensional visual features;

[0106] The specific steps are as follows:

[0107] 1. Image grayscale conversion:

[0108] The high-definition image of each fixed sampling point i in the high-definition image data. Convert to grayscale image This simplifies subsequent edge detection processing.

[0109] 2. Edge detection:

[0110] Grayscale images are processed using the Canny edge detection algorithm. High and low thresholds of 50 and 150 are set respectively to ensure the accuracy and continuity of edges, generating an edge image with a fixed sampling point i. .

[0111] 3. Feature extraction:

[0112] For edge images The Hough Transform is applied to identify the characteristics of straight and curved cracks.

[0113] Record the initiation point, termination point, and geometric characteristics of fatigue cracks, including length and angle.

[0114] 4. Crack marking:

[0115] In the original high-definition image Based on the identified crack features, the location of each fatigue crack is marked, forming an image containing two-dimensional visual features. .

[0116] 1.14) Using preprocessed ultrasonic echo data as a depth profile, we can identify potential crack propagation trends in images and supplement local areas where visual data is limited.

[0117] The specific steps are as follows:

[0118] 1. Depth profile data mapping:

[0119] Basic defect distribution map The depth information in the image is mapped onto a two-dimensional image plane to form a depth profile. .

[0120] 2. Crack propagation trend identification:

[0121] Depth profiles are identified using gradient-based analysis methods. Potential propagation direction and trend of medium cracks.

[0122] Set the extended trend threshold to 40% of the gradient change rate to filter out noise effects.

[0123] 3. Trend Supplement:

[0124] For images Areas with visual blind spots, according to depth profile maps The expansion trend in the crack is analyzed to supplement the potential extension portion of the crack and generate a supplementary distribution map. .

[0125] 1.15) Combining the two-dimensional visual features of high-definition image data with the depth information of ultrasonic data, a preliminary defect distribution map with complete morphology and internal structure is generated;

[0126] The specific steps are as follows:

[0127] 1. Image registration:

[0128] This embodiment employs a feature-point-based registration algorithm, specifically the SIFT algorithm, for images containing two-dimensional visual features. With supplementary distribution map Perform precise alignment to generate a registered image.

[0129] 2. Multimodal fusion:

[0130] Apply a weighted fusion method to register the images and supplementary distribution map The data is fused according to its weight (70% for visual data and 30% for depth information) to generate a fusion defect distribution map. ;

[0131] Distribution map of fusion defects Morphological processing was applied, using expansion and etching operations to further improve the connectivity of the defect edges, ensuring the integrity of the crack morphology, and obtaining the final preliminary defect distribution map. .

[0132] 1.2) Extract defect features from the preliminary defect distribution map of each fixed sampling point to obtain a binarized defect image; the defect features include surface cracks and internal cracks;

[0133] Internal cracks refer to tiny cracks or fracture surfaces inside the test piece, which are caused by the accumulation of stress, fatigue, and other factors during the use of the test piece.

[0134] Defect feature extraction includes region segmentation, using a region-growing-based image segmentation algorithm to separate regions containing surface cracks and internal cracks from the background, forming a binarized defect image;

[0135] Morphological processing is performed on the segmented defect images, using opening and closing operations to remove isolated noise points and fill crack areas, ensuring the continuity and integrity of defect edges.

[0136] 1.3) Match the high-definition image data and defect images acquired at each fixed sampling point according to the fixed sampling point, and upload them to the central database;

[0137] Defect feature information of surface cracks and internal cracks is marked on each fixed sampling point i in the defect image to form initial material fatigue image data with defect feature information;

[0138] The defect feature information includes crack location, crack length, crack width, crack propagation rate, and crack distribution density;

[0139] 1.4) The defect feature information contained in the initial material fatigue image data of each measurement is weighted and averaged to generate uniform material fatigue image data with fixed sampling point i, and the defect feature information in the material fatigue image data is sequentially recorded as crack length. Crack width Crack propagation rate and crack distribution density .

[0140] The specific steps are as follows:

[0141] Using image analysis software, the length (unit: micrometer) and width (unit: micrometer) of each identified fatigue crack at a fixed sampling point i are measured.

[0142] Record the coordinates of the start and end points of each fatigue crack to calculate its geometric characteristics.

[0143] Explanation of crack propagation rate calculation:

[0144] Based on crack length data obtained from multiple ultrasonic echo measurements at different time points, the crack propagation rate (unit: micrometers / minute) is calculated. Specifically, based on historical data and ultrasonic detection auxiliary results, the crack propagation rate within the same fixed sampling point is determined. This value is estimated from the difference in crack length between consecutive frames of images.

[0145] The crack length variation curve over time was fitted using linear regression, and the slope was taken as the crack propagation rate.

[0146] Explanation of the calculation of crack distribution density:

[0147] The unit area is set to 1 square centimeter in this embodiment. The number of cracks is counted in the image area of ​​each fixed sampling point i, and the crack distribution density is calculated (unit: number of cracks / unit area).

[0148] The measured crack parameters are compared with the generated preliminary defect distribution map. The data are fused to form material fatigue image data at a fixed sampling point i;

[0149] The defect feature information contained in the initial material fatigue image data of each measurement is weighted and averaged to generate uniform material fatigue image data. .

[0150] 1.5) Format and store the unified material fatigue image data; the specific logic is as follows:

[0151] The uniform material fatigue image data is saved in a predetermined standard format, TIFF or PNG, to ensure image clarity and data integrity.

[0152] The formatted material fatigue image data is associated with the corresponding fixed sampling point number i and archived, then uploaded to the central database to ensure the convenience of subsequent data processing and analysis.

[0153] 1.6) By comparing the high-definition image data in step S1 with the material fatigue image data, the accuracy and consistency of the defect information generated by the ultrasonic echo data are verified, ensuring the reliability of the material fatigue image data.

[0154] If inconsistencies or errors are found, return to steps 1.1) and 1.2) to re-perform data preprocessing and defect feature extraction until material fatigue image data that meets the requirements is generated.

[0155] Example 2:

[0156] Step S2: Extract the defect feature information of the material fatigue image data from each fixed sampling point to form a fatigue feature parameter set;

[0157] Furthermore, by applying cubic spline interpolation, the fatigue characteristic parameter set is smoothed to form continuous interpolation curve data, thereby obtaining continuous fatigue characteristic interpolation data.

[0158] Further explanation: For the detection area of ​​the stress zone on the surface of the workpiece, a fixed spacing is preset. Arrange N fixed sampling points, where the index of the fixed sampling point i ∈ {1, 2, ..., N}; the spatial position of the fixed sampling point is represented as:

[0159]

[0160] Crack length for obtaining defect feature information corresponding to fixed sampling point i Crack width Crack propagation rate and crack distribution density ;

[0161] The defect feature information of all sampling points is serialized to form a fatigue feature parameter set:

[0162]

[0163] The obtained fatigue characteristic parameter set Interpolation processing is performed using the spatial location of fixed sampling points as the abscissa. Cubic spline interpolation is applied to crack length, crack width, crack propagation rate, and crack density to obtain continuous and smooth interpolation curve data. The interpolation curve data includes the crack length interpolation curve. Crack width interpolation curve Crack propagation rate interpolation curve Crack distribution density interpolation curve ;

[0164] The specific steps are as follows:

[0165] 2.1) Construction of interpolation nodes:

[0166] Set the number of fixed interpolation nodes k=10, evenly distributed within the detection area, and set the index of each interpolation node as j. The positions of the interpolation nodes are denoted as:

[0167]

[0168] 2.2) Construction of the cubic spline interpolation model:

[0169] Crack length parameter sequence Construct a cubic spline interpolation model, and the interpolation curve is represented as:

[0170]

[0171] in, Let f be a basis function, and let j be the cubic spline function. The weights of interpolation node j are used, and the objective is to minimize the fitting error.

[0172]

[0173] Using the same method, interpolation models were constructed for crack width, crack propagation rate, and crack distribution density, and their interpolation curves are denoted as follows:

[0174]

[0175]

[0176]

[0177] Weighting coefficient , , Calculated using standard least squares method to minimize interpolation error;

[0178] 2.3) Generation of interpolation curves:

[0179] Four continuous interpolation curves were calculated using the cubic spline interpolation formula: crack length interpolation curve. Crack width interpolation curve Crack propagation rate interpolation curve Crack distribution density interpolation curve ;

[0180] Composite crack length interpolation curve Crack width interpolation curve Crack propagation rate interpolation curve Crack distribution density interpolation curve To form a complete set of continuous fatigue feature interpolation data using a unified representation, and each set of interpolation data is denoted as . .

[0181] The specific steps are as follows:

[0182] 2.4) Unified representation of datasets:

[0183] The generated crack length interpolation curve Crack width interpolation curve Crack propagation rate interpolation curve Crack distribution density interpolation curve The data is uniformly categorized and formed into an interpolated data set F(x), whose expression is:

[0184]

[0185] The spatial range of the continuous interpolation dataset F(x) is defined as the range of the distribution of fixed sampling points within the detection area, i.e.:

[0186]

[0187] 2.5) Data Presentation Output:

[0188] F(x) is used as continuous data of material fatigue characteristics for subsequent assessment of fatigue state and prediction of defect propagation trend.

[0189] Example 3:

[0190] Step S3: Synchronously fuse the continuous fatigue feature interpolation data with the dynamic and static detection data to form a comprehensive dataset;

[0191] definition It is the representation value of the dynamic and static detection data at a fixed sampling point i; and ;

[0192] in, It is the vibration frequency index at a fixed sampling point i; It is the vibration amplitude index at a fixed sampling point i; It is the comprehensive vibration index at a fixed sampling point i; It is the temperature index at a fixed sampling point i; It is the temperature change rate index at a fixed sampling point i; It is the comprehensive temperature index at a fixed sampling point i;

[0193] definition Let represent the set of interpolated data for a fixed sampling point i within the detection area, and ;

[0194] in, It is an index marker for the spatial location of a fixed sampling point i;

[0195] 3.1) Simultaneously collect dynamic and static test data and material fatigue image data at each fixed sampling point within the test area, and add a uniform timestamp to all data; in this embodiment, the time deviation is set to no more than 1 second to ensure the accuracy of time matching of dynamic and static test data; the specific operation steps are as follows:

[0196] A high-precision synchronous acquisition device is used to ensure that dynamic and static test data and material fatigue image data can be recorded simultaneously.

[0197] It should be noted that a high-precision synchronous acquisition device refers to a hardware system with synchronous data acquisition capabilities. It supports the joint acquisition of multiple data sources, records the timestamp of each sampling point, and strictly synchronizes each acquisition channel using a unified clock, with the maximum allowable time deviation not exceeding a preset time limit. The specific working principle is as follows:

[0198] Multi-channel data acquisition: The high-precision synchronous acquisition device is connected to each sensor in the detection area, and simultaneously activates multiple acquisition modules to sample vibration, temperature and image data at fixed time intervals.

[0199] Time Synchronization: The high-precision clock inside the high-precision synchronous acquisition device provides a unified time reference for all acquisition modules. At each data acquisition moment, the device automatically generates a timestamp for each set of data and uses an error correction module to adjust for time errors between different data channels, ensuring the consistency of all acquired data in the time dimension.

[0200] Data Integration and Output: The collected vibration, temperature, and image data are integrated in real time through an embedded data processing module to generate complete data records. After local storage, the data is uploaded to a server or database system via a transmission interface for further analysis.

[0201] Timestamp matching settings: Configure the acquisition device to ensure that the timestamp error between dynamic and static test data and material fatigue image data does not exceed [the specified value]. Second.

[0202] The synchronous acquisition device is activated to collect dynamic and static test data and material fatigue image data from all fixed sampling points, and the timestamp of each set of data is recorded. And a fixed sampling point index i. After the acquisition is completed, the set of dynamic and static detection data of N fixed sampling points is represented as:

[0203]

[0204] The set of continuous fatigue feature interpolation data with N fixed sampling points is represented as:

[0205]

[0206] 3.2) Based on the timestamp of a fixed sampling point i and spatial location For dynamic and static detection data sets and continuous fatigue feature interpolation data set Perform point-by-point matching, and merge the matched data with fixed sampling point numbers to form a comprehensive dataset. ;in It is the composite data of a fixed sampling point i; It is a representation of a comprehensive dataset.

[0207] The specific steps are as follows:

[0208] For each fixed sampling point i∈{1,2,…,N}, according to its timestamp and spatial location , dynamic and static detection data Interpolation data with continuous fatigue characteristics Achieve point-to-point matching;

[0209] Dynamic and static detection data at each fixed sampling point and fatigue feature interpolation data The merger will result in a comprehensive data record. :

[0210]

[0211] By aggregating the comprehensive data records from all fixed sampling points, a comprehensive dataset is formed:

[0212]

[0213] A weighted average method is used to fuse dynamic and static test data and fatigue feature interpolation data, with weights set. and satisfy The comprehensive formula is:

[0214]

[0215] Choose MySQL as the database and design the table structure, including the following fields: .

[0216] Comprehensive dataset Import the data into the database according to the designed field format, and verify the integrity and accuracy of the data.

[0217] Example 4:

[0218] Step S4: Receive the comprehensive dataset and construct a fault prediction model based on continuous fatigue feature interpolation data and dynamic and static detection data;

[0219] Further explanation: 4.1) Select continuous fatigue feature interpolation data from the comprehensive dataset D. With dynamic and static detection data A multivariate nonlinear functional relationship is established to obtain a fault prediction model for the device under test; and the output of the fault prediction model is represented as a fault risk index with a value range of (0,1). ;

[0220] Fault prediction model The following characterization is performed:

[0221]

[0222] Among them, the function This represents a multivariate nonlinear relationship; this embodiment can employ nonlinear modeling methods such as neural networks and support vector machines.

[0223] A multilayer feedforward neural network is selected as the function. Specific form;

[0224] For multilayer feedforward neural networks, it should be noted that:

[0225] The input layer of a multilayer feedforward neural network has 6 nodes, corresponding to the input features;

[0226] Design two hidden layers, each containing 10 neurons, with fully connected layers between them;

[0227] Each hidden layer uses the ReLU (Rectified Linear Unit) activation function, and the output layer uses a linear activation function to suit the needs of regression tasks.

[0228] The output layer contains one neuron and outputs a fault risk index. The value range is within (0,1);

[0229] To train and validate a neural network model, prepare your data by following these steps:

[0230] Dataset splitting: The comprehensive dataset D is divided into a training set (80%), a validation set (10%), and a test set (10%) to ensure that each subset is representative and balanced.

[0231] Data normalization: Standardize the input features so that their mean is 0 and their standard deviation is 1, in order to speed up model training and improve training stability.

[0232] Mean Squared Error (MSE) is used as the loss function;

[0233] The Adam optimizer was chosen for parameter updates because it has good performance and convergence speed when dealing with non-convex optimization problems.

[0234] Training parameter settings:

[0235] Learning Rate: Set to 0.001.

[0236] Batch Size: Set to 32.

[0237] Training epochs: Set to 1000, or until the validation set is complete. .

[0238] An early stopping mechanism is implemented: training is stopped when the loss on the validation set does not show a significant decrease within 50 consecutive rounds to prevent overfitting.

[0239] 4.2) The least squares method is used to determine the weights and biases of the neural network in the fault prediction model by fitting historical fault data; the convergence criterion for the parameters is set as the coefficient of determination during the fitting process. ;

[0240] The specific steps are as follows:

[0241] Collect and organize a comprehensive historical dataset containing known fault states. To ensure data quality and representativeness;

[0242] The weights and biases of the neural network in the fault prediction model are initialized; in this embodiment, small random values ​​are used.

[0243] Define the objective function as the sum of squared errors between the predicted values ​​and the actual fault conditions:

[0244]

[0245] in, This represents the actual fault state of the h-th historical sample.

[0246] The model parameters are optimized using a fixed least squares method, iteratively adjusting the parameters to minimize the objective function; the optimization formula is as follows:

[0247]

[0248] in, This represents the set of model parameters.

[0249] Monitoring the coefficient of determination during the iteration process The calculation formula is as follows:

[0250]

[0251] when When the iteration stops, it is considered that the model parameters have converged.

[0252] The predictive performance of the fault prediction model was validated using an independent validation set to ensure the model's generalization ability.

[0253] 4.3) The constructed fault prediction model is applied to real-time data detection. The system updates the model input data every 5 minutes to achieve continuous monitoring of the test component corresponding to the detection object.

[0254] The specific steps are as follows:

[0255] Real-time data acquisition: Collecting the latest comprehensive data sets in real time through sensors and data acquisition systems. .

[0256] The real-time collected data undergoes the same preprocessing steps as historical data, including normalization and numerical mapping, to ensure that the data is consistent with the model input.

[0257] The preprocessed real-time data is input into the fault prediction model to calculate the current fault risk index. ;

[0258] Configure a system scheduled task to automatically perform data acquisition, preprocessing, and model calculation steps every 5 minutes to ensure the current fault risk index is maintained. Updated in real time.

[0259] If the current failure risk index The closer the value of is to 1, the higher the failure risk of the device under test;

[0260] Set failure risk index The threshold is ; The determination was made by an expert panel using fuzzy hierarchical analysis (FAHP) based on experimental data; specifically, it was executed using Python or MATLAB.

[0261] When the failure risk index Exceeding the predetermined threshold When this occurs, an alarm mechanism is triggered, indicating that the device under test has a potential failure risk.

[0262] 4.3) The constructed fault prediction model and its data processing flow are implemented in an embedded processor to ensure the continuity of the data acquisition, preprocessing, interpolation, fusion and model calculation process, meet the real-time detection requirements, and update all data indicators with fixed values ​​as output results.

[0263] Embedded processor selection and configuration: Select a high-performance embedded processor (such as the ARM Cortex series) as the system core, and configure sufficient memory and storage space to support real-time data processing and model calculation.

[0264] The parameters and structure of the fault prediction model are deployed to an embedded processor, and the efficient computing library TensorFlow Lite is used to optimize the model's running efficiency.

[0265] Develop a data processing module to implement the following functions:

[0266] The data acquisition interface communicates with sensors and image acquisition devices to obtain real-time data.

[0267] The data preprocessing process includes normalization and numerical mapping.

[0268] The fusion processing of interpolated data and integrated dynamic and static detection data.

[0269] The model calculation module takes the processed data as input and outputs the fault risk index.

[0270] All modules were integrated into the embedded system, and overall system testing was conducted to ensure the consistency of each process module and the accuracy of the data. The testing included:

[0271] Configure the output interface to display the fault risk index. Output the data to the monitoring platform or alarm system in a fixed numerical format to ensure real-time updates and easy reading.

[0272] Example 5:

[0273] Step S5: Obtain the fatigue characteristic parameter set and perform comprehensive weighted calculation and analysis to construct a defect propagation risk index. The defect propagation risk index is used to assess the crack propagation risk level of the test piece.

[0274] Further explanation: Crack length at fixed sampling point i Crack width Crack propagation rate and crack distribution density A defect expansion risk index for a fixed sampling point i is constructed by weighted normalization. Defect propagation risk indicators Used to quantify the risk level of crack propagation at a fixed sampling point i;

[0275] Defect Expansion Risk Indicators The calculation formula is as follows:

[0276]

[0277]

[0278] in, It is a defect propagation risk index for a fixed sampling point i, ranging between (0,1); It is the normalized crack length at a fixed sampling point i; It is the crack width at the fixed sampling point i after normalization; It is the normalized crack propagation rate at a fixed sampling point i; It is the normalized crack distribution density at a fixed sampling point i; ;

[0279] , These are the crack lengths at all fixed sampling points. Crack width Crack propagation rate and crack distribution density The minimum and maximum values;

[0280] They are The weighting coefficients satisfy .

[0281] Based on engineering experience or statistical analysis, assign weights to each parameter. This ensures that their sum is 1. This example provides:

[0282]

[0283] set up The value range is the interval (0,1); the crack propagation risk level assessment results are as follows:

[0284] when The closer it is to 0, the more it indicates The lower the value, the lower the fatigue risk level of the fixed sampling point i on the test piece;

[0285] when The closer it is to 1, the more it indicates The higher the value, the higher the fatigue risk level of the fixed sampling point i on the test piece;

[0286] set up warning threshold ,and Values ​​within range Internal selection; this embodiment The value is 0.65; and The determination was made by an expert panel using fuzzy hierarchical analysis (FAHP) based on experimental data; specifically, it was executed using Python or MATLAB.

[0287] In N fixed sampling points, when more than half of the fixed sampling points exist... When the value is 0, it indicates that the fatigue risk level of the test piece is high.

[0288] It should be noted that: The logical reasoning behind the changes in formula parameters is as follows:

[0289] How do changes in the parameters in the formula affect value:

[0290] Crack length Increase, Increase;

[0291] Technical consequences: Increased fatigue risk and increased crack propagation in materials;

[0292] Crack width Increase, Increase;

[0293] Technical consequences: Increased fatigue risk and increased crack propagation depth.

[0294] Crack propagation rate Increase, Increase;

[0295] Technical consequences: Increased fatigue risk and accelerated crack propagation.

[0296] Crack distribution density Increase, Increase;

[0297] Technical consequences: Increased fatigue risk and more dense distribution of cracks throughout the material.

[0298] Example 6:

[0299] Step S6: Obtain the local stress values ​​and overall stress values ​​at multiple fixed sampling points on the test piece, determine the stress concentration of the test piece by calculating the ratio, and obtain the potential risk assessment results for each fixed sampling point;

[0300] Further explanation: The stress concentration of the test piece is determined by ratio calculation, and the potential risk assessment results of each fixed sampling point are obtained, specifically including;

[0301] Let i1 be the index of a fixed sampling point with a crack, and i1∈{1,2,…,N1}, N1≤N. Collect the local stress in the crack region at the fixed sampling point i1. Overall stress at fixed sampling point i1 and local stress With overall stress Stress concentration criteria values ​​in the range (0,1) are constructed using a ratio method. To quantify the stress concentration at a fixed sampling point i1;

[0302] The calculation formula is:

[0303]

[0304] in, It is the stress concentration determination value of the fixed sampling point i1, which is in the range of (0,1);

[0305] It is the local stress (MPa) in the crack region at a fixed sampling point i1.

[0306] It is the overall stress (MPa) at a fixed sampling point i1;

[0307] It should be noted that: overall stress The stress value refers to the stress value of the material under test under overall stress, reflecting the stress state of the material in the entire structure.

[0308] Local stress The stress value of the material under test at the crack region at a fixed sampling point i1 reflects the stress concentration at that location.

[0309] The selection of the crack region needs to effectively reflect the stress concentration condition; the potential risk assessment results are as follows:

[0310] when The closer it is to 0, the greater the local stress. The lower than the overall stress The more uniform the stress distribution at the fixed sampling point i1, the lower the stress concentration, and the better the health status of the test piece in the area of ​​the fixed sampling point i1.

[0311] when The closer a value is to 1, the stronger the local stress. The closer it gets to the overall stress The more concentrated the stress is at the fixed sampling point i1, the higher the stress near the crack at the fixed sampling point i1, and the greater the potential risk of fatigue propagation or fracture.

[0312] set up warning threshold ,and Values ​​within range Internal selection; this embodiment The value is 0.71; and The determination was made by an expert panel using fuzzy hierarchical analysis (FAHP) based on experimental data; specifically, it was executed using Python or MATLAB.

[0313] In N1 fixed sampling points containing cracks, when more than half of the fixed sampling points... When the test piece is in a high-risk state, it indicates that the potential fatigue propagation or fracture risk is high.

[0314] It is necessary to strengthen monitoring or take measures to alleviate stress concentration and prevent the deterioration of the material properties of the test piece;

[0315] It should be noted that: The logical reasoning behind the changes in formula parameters is as follows:

[0316] Local stress Increase, Increase;

[0317] Technical consequences: Increased stress concentration leads to increased stress values ​​near cracks, posing a risk of fatigue propagation or fracture.

[0318] Overall stress Increase, Decrease;

[0319] Technical benefits: Reduced stress concentration, more uniform stress distribution in materials, and reduced fatigue risk.

[0320] Example 7:

[0321] Step S7: Based on the crack propagation risk level assessment results and potential risk assessment results of the test piece, provide correction strategies for the output results of the fault prediction model, and apply the optimized output results to the fault prediction comparison process of the test piece in the next testing period.

[0322] Further explanation: When the fatigue risk level assessment result of the crack propagation risk level of the test piece is determined to be high; and the potential fatigue propagation or fracture risk of the potential risk assessment result is also high, a predetermined threshold will be applied. Implement the first-level adjustment strategy;

[0323] The primary adjustment strategy, specifically the adjustment logic, includes:

[0324]

[0325] in, This is the adjusted predetermined threshold. It is the first adjustment factor, and The value is within the range (0.01, 0.25);

[0326] It should be noted that, The setting is used to set a predetermined threshold. To reduce the risk, since both the crack propagation risk assessment and the potential risk assessment results are rated as high, the failure risk of the test part is considered high. Therefore, the predetermined threshold needs to be lowered. By lowering the predetermined threshold, the alarm mechanism can be triggered earlier, making the system more sensitive to high-risk states and taking swift action; this is a protective measure to prevent further damage to the system from faults at high-risk levels.

[0327] When only one of the crack propagation risk level assessment results and the potential risk assessment results of the test piece is classified as high, a predetermined threshold is applied. Implement a level-two adjustment strategy;

[0328] The secondary adjustment policy, specifically the adjustment logic, includes:

[0329]

[0330] in, This is the adjusted predetermined threshold. It is the second adjustment factor, and The value is within the range (0.01, 0.25). This design aims to make the adjustment range of the primary adjustment strategy greater than that of the secondary adjustment strategy.

[0331] When neither the crack propagation risk level assessment nor the potential risk assessment result of the test piece is classified as high-level, no predetermined threshold is applied. Implement adjustment strategies;

[0332] The adjustment range of the first-level adjustment strategy is greater than that of the second-level adjustment strategy.

[0333] It should be noted that all calculation formulas in this application employ regression analysis, including but not limited to machine learning algorithms, to deeply analyze the collected parameters and identify their natural trends and interrelationships. Specialized software, such as Python's Scikit-learn library or the R language, is used to automatically generate mathematical models that match the data. Then, cross-validation and other methods are used to objectively evaluate the model performance, and continuous feedback and optimization are combined to ensure that the created formulas truly reflect the inherent laws of the data, thereby guaranteeing their effectiveness and accuracy. In all calculation formulas in this application, the parameters in each formula undergo dimensionless processing within a consistent range to ensure that different physical quantities are compared on the same scale; dimensionless processing techniques include, but are not limited to, Min-Max Normalization and Z-Score standardization.

[0334] The technical solution of this invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of this invention.

[0335] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0336] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0337] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data, characterized in that, The specific steps include: Step S1: Collect high-definition image data and ultrasonic echo data of multiple fixed sampling points on the surface of the test piece under a fixed environment, and analyze these data to generate material fatigue image data for each fixed sampling point; at the same time, collect dynamic and static test data of the test piece. Step S2: Extract the defect feature information of the material fatigue image data from each fixed sampling point to form a fatigue feature parameter set; Furthermore, by applying cubic spline interpolation, the fatigue characteristic parameter set is smoothed to form continuous interpolation curve data, thereby obtaining continuous fatigue characteristic interpolation data. For the detection area of ​​the stress region on the surface of the workpiece, a fixed spacing is preset. Arrange N fixed sampling points, where the index of the fixed sampling point i ∈ {1, 2, ..., N}; the spatial position of the fixed sampling point is represented as: ； Crack length for obtaining defect feature information corresponding to fixed sampling point i Crack width Crack propagation rate and crack distribution density ; The defect feature information of all sampling points is serialized to form a fatigue feature parameter set: ； The obtained fatigue characteristic parameter set Interpolation processing is performed using the spatial location of fixed sampling points as the abscissa. Cubic spline interpolation is applied to crack length, crack width, crack propagation rate, and crack density to obtain continuous and smooth interpolation curve data. The interpolation curve data includes the crack length interpolation curve. Crack width interpolation curve Crack propagation rate interpolation curve Crack distribution density interpolation curve ; Composite crack length interpolation curve Crack width interpolation curve Crack propagation rate interpolation curve Crack distribution density interpolation curve To form a complete set of continuous fatigue feature interpolation data using a unified representation, and each set of interpolation data is denoted as . ; Step S3: Synchronously fuse the continuous fatigue feature interpolation data with the dynamic and static detection data to form a comprehensive dataset; Step S4: Receive the comprehensive dataset and construct a fault prediction model based on continuous fatigue feature interpolation data and dynamic and static detection data; Step S5: Obtain the fatigue characteristic parameter set and perform comprehensive weighted calculation and analysis to construct a defect propagation risk index. The defect propagation risk index is used to assess the crack propagation risk level of the test piece. Step S6: Obtain the local stress values ​​and overall stress values ​​at multiple fixed sampling points on the test piece, determine the stress concentration of the test piece by calculating the ratio, and obtain the potential risk assessment results for each fixed sampling point; Step S7: Based on the crack propagation risk level assessment results and potential risk assessment results of the test piece, provide correction strategies for the output results of the fault prediction model, and apply the optimized output results to the fault prediction comparison process of the test piece in the next testing period.

2. The comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data according to claim 1, characterized in that: The process involves acquiring continuous high-definition image data and ultrasonic echo data of the surface of the test piece under a fixed environment, as well as acquiring dynamic and static test data of the test piece, specifically including: Under fixed conditions, the stress-bearing area on the surface of the workpiece to be tested is selected as the detection area; Images are acquired from N fixed sampling points distributed across the detection area to obtain high-resolution image data for each fixed sampling point; the index of each fixed sampling point is denoted as i, where i∈{1,2,…,N}; where i represents the index of the fixed sampling point. Acquire ultrasonic echo data from at least three repeated measurements at each fixed sampling point; Dynamic and static test data include vibration and temperature parameters of the test piece; Vibration parameters include vibration frequency index. and vibration amplitude index ; Define the vibration frequency index By measuring the vibration frequency of the test component during operation and normalizing it to reflect its state within a safe range; this vibration frequency index... The closer it is to 1, the higher the vibration frequency, which indicates a greater risk of mechanical failure or wear of the test piece. Define vibration amplitude index The severity of vibration of the test piece is assessed by measuring and normalizing the amplitude of piston vibration; the vibration amplitude index. The higher the value, the greater the risk of mechanical failure or wear of the part under test; By using a weighted average method, combined with the vibration frequency index and vibration amplitude index Generate a comprehensive vibration index ; The range is within the interval (0,1); The closer the value is to 1, the higher the risk of failure of the device under test. Setting temperature parameters includes temperature index and temperature change rate index ; Temperature Index By measuring and normalizing the temperature of the test piece, the impact of temperature conditions on thermal fatigue and overheating of the test piece material is assessed; temperature index. The higher the value, the higher the temperature, and the greater the potential risk of thermal fatigue and overheating; Temperature change rate index The impact of temperature change rate on thermal shock and structural damage of the test specimen is assessed by measuring and normalizing the rate of temperature change over a period of time; the temperature change rate index. The higher the value, the faster the temperature changes and the greater the potential risk of thermal shock. By using a weighted average method, combined with the temperature index and temperature change rate index Generate a comprehensive temperature index The comprehensive temperature index The value range is within the interval (0,1); comprehensive temperature index The closer the value is to 1, the higher the degree of thermal fatigue of the test piece.

3. The comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data according to claim 2, characterized in that: Generating material fatigue image data for each fixed sampling point specifically includes: The high-definition image data of each fixed sampling point and the ultrasonic echo data of at least three repeated measurements are preprocessed, and a preliminary defect distribution map of each fixed sampling point is generated based on each repeated measurement. Defect features are extracted from the preliminary defect distribution map at each fixed sampling point to obtain a binarized defect image; the defect features include surface cracks and internal cracks. The high-definition image data and defect images acquired at each fixed sampling point are matched according to the fixed sampling points and uploaded to the central database; Defect feature information of surface cracks and internal cracks is marked on each fixed sampling point i in the defect image to form initial material fatigue image data with defect feature information; The defect feature information includes crack location, crack length, crack width, crack propagation rate, and crack distribution density; The defect feature information contained in the initial material fatigue image data of each measurement is weighted and averaged to generate uniform material fatigue image data at a fixed sampling point i. The defect feature information in the material fatigue image data is then sequentially recorded as crack length. Crack width Crack propagation rate and crack distribution density .

4. The comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data according to claim 3, characterized in that: definition It is the representation value of the dynamic and static detection data at a fixed sampling point i; and ; definition Let represent the set of interpolated data for a fixed sampling point i within the detection area, and ; Dynamic and static test data and material fatigue image data are collected synchronously at each fixed sampling point within the test area, and a unified timestamp is added to all data. The set of dynamic and static detection data with N fixed sampling points is represented as: ； The set of continuous fatigue feature interpolation data with N fixed sampling points is represented as: ； Based on the timestamp of fixed sampling point i and spatial location For dynamic and static detection data sets and continuous fatigue feature interpolation data set Perform point-by-point matching, and merge the matched data with fixed sampling point numbers to form a comprehensive dataset. ;in It is the composite data of a fixed sampling point i; It is a representation of a comprehensive dataset.

5. The comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data according to claim 4, characterized in that: Select continuous fatigue feature interpolation data from the comprehensive dataset D. With dynamic and static detection data A multivariate nonlinear functional relationship is established to obtain a fault prediction model for the device under test; and the output of the fault prediction model is represented as a fault risk index with a value range of (0,1). ; The least squares method is used to determine the weights and biases of the neural network in the fault prediction model by fitting historical fault data; the coefficient of determination is set as the convergence criterion for the parameters during the fitting process. ; Calculate the current failure risk index ; If the current failure risk index The closer the value of is to 1, the higher the failure risk of the device under test; Set failure risk index The threshold is ; When the failure risk index Exceeding the predetermined threshold When this occurs, an alarm mechanism is triggered, indicating that the device under test has a potential failure risk.

6. The comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data according to claim 5, characterized in that: Crack length at a fixed sampling point i Crack width Crack propagation rate and crack distribution density A defect expansion risk index for a fixed sampling point i is constructed by weighted normalization. Defect propagation risk indicators Used to quantify the risk level of crack propagation at a fixed sampling point i; set up The value range is the interval (0,1); the crack propagation risk level assessment results are as follows: when The closer it is to 0, the more it indicates The lower the value, the lower the fatigue risk level of the fixed sampling point i on the test piece; when The closer it is to 1, the more it indicates The higher the value, the higher the fatigue risk level of the fixed sampling point i on the test piece; set up warning threshold ,and Values ​​within range Internal selection; In N fixed sampling points, when more than half of the fixed sampling points exist... When the value is 0, it indicates that the fatigue risk level of the test piece is high.

7. The comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data according to claim 6, characterized in that: The stress concentration of the test piece is determined by ratio calculation, and the potential risk assessment results of each fixed sampling point are obtained, including: Let i1 be the index of a fixed sampling point with a crack, and i1∈{1,2,…,N1}, N1≤N. Collect the local stress in the crack region at the fixed sampling point i1. Overall stress at fixed sampling point i1 and local stress With overall stress Stress concentration criteria values ​​in the range (0,1) are constructed using a ratio method. To quantify the stress concentration at a fixed sampling point i1; The results of the potential risk assessment are as follows: when The closer it is to 0, the greater the local stress. The lower than the overall stress The more uniform the stress distribution at the fixed sampling point i1, the lower the stress concentration, and the better the health status of the test piece in the area of ​​the fixed sampling point i1. when The closer a value is to 1, the stronger the local stress. The closer it gets to the overall stress ; The more concentrated the stress is at the fixed sampling point i1, the higher the stress near the crack at the fixed sampling point i1, and the greater the potential risk of fatigue propagation or fracture. set up warning threshold ,and Values ​​within range Internal selection; In N1 fixed sampling points containing cracks, when more than half of the fixed sampling points... When the value is 0, it indicates that the potential fatigue propagation or fracture risk of the test piece is at a high level.

8. The comparative analysis and optimization method based on cubic spline interpolation dynamic and static detection data according to claim 7, characterized in that: When the fatigue risk level assessment result of the crack propagation risk level of the test piece is determined to be high; and the potential fatigue propagation or fracture risk of the potential risk assessment result is also high, a predetermined threshold is applied. Implement the first-level adjustment strategy; When only one of the crack propagation risk level assessment results and the potential risk assessment results of the test piece is classified as high, a predetermined threshold is applied. Implement a level-two adjustment strategy; When neither the crack propagation risk level assessment nor the potential risk assessment result of the test piece is classified as high-level, no predetermined threshold is applied. Implement adjustment strategies; The adjustment range of the first-level adjustment strategy is greater than that of the second-level adjustment strategy.

Images