A method for predicting spectral temperature luminance degradation of a polyurethane functional material component
By constructing a weighted joint regression network model, the applicability of degradation assessment for polyurethane functional material components was addressed, enabling accurate prediction of their performance changes and supporting their rational use and maintenance.
Patent Information
- Application Number
- CN202310710487.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-15
- Publication Date
- 2026-03-03
- Estimated Expiration
- 2043-06-15
AI Technical Summary
The lack of a widely applicable and scalable degradation assessment system for polyurethane functional material components in the current technology has resulted in insufficient research on the changes in their performance over time, affecting their use and maintenance.
A weighted joint regression network prediction model is constructed. By using the detection results of spectral, temperature, and brightness feature images for weighted joint regression, the degradation of polyurethane functional material components is predicted, avoiding local optima and overfitting problems.
This enables more accurate prediction of the performance changes of polyurethane functional material components, supporting their rational use and maintenance.
Smart Images
Figure CN116994678B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of materials testing technology, and in particular to a method for predicting the spectral temperature brightness degradation of polyurethane functional material components. Background Technology
[0002] Based on foreign fundamental theories, a series of functional material degradation assessment methods have been established in China. However, the systematic nature and applicability of these assessments still lag behind those abroad, and a highly applicable and widely applicable assessment system has not yet been formed.
[0003] Polyurethane functional materials are commonly used materials, but there is very little research on how their properties change over time, which is not conducive to the use, replacement and maintenance of polyurethane functional material components. Summary of the Invention
[0004] This invention discloses a method for predicting the spectral temperature brightness degradation of polyurethane functional material components. It can fit the law of change of the performance of polyurethane functional materials over time, which is beneficial to the use, replacement and maintenance of polyurethane functional materials.
[0005] It is achieved through the following technical solution:
[0006] A weighted joint regression network prediction model was constructed to predict polyurethane functional material components.
[0007] The spectral similarity, temperature similarity, and brightness similarity of polyurethane functional material components were predicted using a weighted joint regression network prediction model.
[0008] The advantage of this embodiment is that it proposes a degradation characterization model for the temperature, brightness, and spectral features of functional materials based on a weighted joint regression network. By using weighted joint regression of temperature, brightness, and spectral feature image detection results, compared with a traditional single regression fitting algorithm, it can avoid getting trapped in local optimum and overfitting, making the prediction results closer to the reality.
[0009] Furthermore, the specific method for constructing the weighted joint regression network prediction model is as follows:
[0010] Select test components that are identical to the polyurethane functional material components to be predicted;
[0011] Spectral images, temperature images, and brightness images of the test components at different test periods were obtained through degradation tests.
[0012] Process spectral test images, temperature test images, and brightness test images, obtain and label the spectral similarity, temperature similarity, and brightness similarity of different test periods, and construct training samples;
[0013] A weighted joint regression network prediction model is trained using training samples.
[0014] Furthermore, the specific method for training the weighted joint regression network prediction model using training samples is as follows:
[0015] The training data is divided into first training data and second training data;
[0016] The least squares regression model, lasso regression model, and ridge regression model are trained using the first training data. These models are then substituted into the weighted joint regression network prediction model, and the formula is as follows:
[0017]
[0018] In the formula, X is the time series dataset of functional material testing results, y is the similarity, and w1 = (X T X) -1 X T y, w2 = (X T X+λI n ) -1 X T y, α are the weight coefficients of the least squares regression model, β are the weight coefficients of lasso regression and ridge regression, and λ is the regularization coefficient;
[0019] Based on the stepwise regression technique of support vector regression, the n coefficients of the weighted joint regression network prediction model are adjusted, n-1 coefficients are fixed, and the remaining coefficient is substituted into the second training data to establish a regression model using the support vector machine regression algorithm, so as to obtain the corresponding model and cost. The coefficient value with the minimum cost is replaced with the original coefficient.
[0020] Repeat the above operation for the remaining n-1 coefficients until all n coefficients are gradually corrected by regression.
[0021] Furthermore, the degradation test involves acquiring spectral test images, temperature test images, and brightness test images of the test component at different times, under the same temperature, the same light, the same natural environment, and the same camera equipment.
[0022] The sampling period is once per quarter, with more than 12 sampling times.
[0023] Furthermore, the specific methods for processing the spectral test image, the temperature test image, and the brightness test image are as follows:
[0024] The spectral test images are processed using a classic anomaly detector to extract spectral feature data;
[0025] The temperature test image and the brightness test image are processed using a classic edge detection operator to extract temperature feature data and brightness feature data.
[0026] A time series is constructed by fusing spectral feature data, temperature feature data, and brightness feature data.
[0027] Furthermore, the time series is labeled with the experimental shooting time to construct training samples.
[0028] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained from the following description and claims. Attached Figure Description
[0029] The accompanying drawings of this invention are described below.
[0030] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation
[0031] The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0032] A method for predicting the spectral temperature brightness degradation of polyurethane functional material components, such as Figure 1 As shown, the specific method is as follows:
[0033] S1. Construct a weighted joint regression network prediction model for predicting polyurethane functional material components.
[0034] Specifically, the method for constructing the weighted joint regression network prediction model is as follows:
[0035] S11. Select a test component that is the same as the polyurethane functional material component to be predicted;
[0036] S12. Obtain spectral test images, temperature test images, and brightness test images of the test components at different test periods through degradation tests;
[0037] Specifically, the degradation test involves acquiring spectral test images, temperature test images, and brightness test images of the test component at different times, under the same temperature, the same light, the same natural environment, and the same camera equipment; the sampling period is once per quarter, and the number of collections is more than 12.
[0038] S13. Process spectral test images, temperature test images, and brightness test images, obtain and label the spectral similarity, temperature similarity, and brightness similarity for different test periods, and construct training samples;
[0039] Specifically, the method for processing the spectral test image, the temperature test image, and the brightness test image is as follows:
[0040] S131. The spectral test image is processed using a classic anomaly detector to extract spectral feature data;
[0041] S132. The temperature test image and the brightness test image are processed using a classic edge detection operator to extract temperature feature data and brightness feature data;
[0042] S133. By fusing spectral feature data, temperature feature data, and brightness feature data, a time series is constructed.
[0043] S14. Train the weighted joint regression network prediction model using the training samples.
[0044] Specifically, the method for training a weighted joint regression network prediction model using training samples is as follows:
[0045] S141. Divide the training data into first training data and second training data;
[0046] S142. Train the least squares regression model, lasso regression model, and ridge regression model using the first training data, and substitute them into the weighted joint regression network prediction model. The formula is as follows:
[0047]
[0048] In the formula, X is the time series dataset of functional material testing results, y is the similarity, and w1 = (X T X) -1 X T y, w2 = (X T X+λI n ) -1 X T y, α are the weight coefficients of the least squares regression model, β are the weight coefficients of lasso regression and ridge regression, and λ is the regularization coefficient;
[0049] S143. Based on the stepwise regression technique of support vector regression, the n coefficients of the weighted joint regression network prediction model are adjusted, n-1 coefficients are fixed, and the remaining coefficient is substituted into the second training data to establish a regression model using the support vector machine regression algorithm, so as to obtain the corresponding model and cost. The coefficient value with the minimum cost is replaced with the original coefficient.
[0050] S144. Repeat the above operation for the remaining n-1 coefficients until all n coefficients are gradually corrected by regression.
[0051] S2. Predict the spectral similarity, temperature similarity, and brightness similarity of polyurethane functional material components using a weighted joint regression network prediction model.
[0052] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0053] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0054] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0055] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.
Claims
1. A method of predicting spectral temperature luminance degradation of a polyurethane functional material member, characterized by, The specific method is as follows: A weighted joint regression network prediction model for predicting a polyurethane functional material component is constructed. The spectral similarity, temperature similarity and brightness similarity of the polyurethane functional material component are predicted by the weighted joint regression network prediction model. The specific method for constructing the weighted joint regression network prediction model is as follows: Select a test component same as the polyurethane functional material component to be predicted; Obtain spectral test images, temperature test images and brightness test images of the test component at different test periods through degradation tests; Process the spectral test images, temperature test images and brightness test images, obtain and label the spectral similarity, temperature similarity and brightness similarity at different test periods, and construct training samples; Train the weighted joint regression network prediction model with the training samples. The specific method for training the weighted joint regression network prediction model with the training samples is as follows: Divide the training data into first training data and second training data; Train a least square regression model, a lasso regression model and a ridge regression model with the first training data, and substitute them into the weighted joint regression network prediction model, which has the following formula: In the formula, X is a functional material detection result time series data set, is a similarity, , , alpha is the weight coefficient of the least squares regression model, beta is the weight coefficient of lasso regression and ridge regression, and lambda is the regularization coefficient. Based on the network stepwise regression technology of support vector regression, adjust the n coefficients of the weighted joint regression network prediction model, fix n-1 coefficients, and use the support vector machine regression algorithm to substitute the remaining one coefficient into the second training data to establish a regression model, obtain the corresponding model and cost, and replace the original coefficient with the coefficient value when the cost is the smallest; Repeat the above operation for the remaining n-1 coefficients until the n coefficients are gradually corrected.
2. The method of claim 1, wherein the polyurethane functional material member is a polyurethane film. The degradation test obtains spectral test images, temperature test images and brightness test images of the test component at different times under the same temperature, same light, same natural environment and same camera equipment. The sampling period is every quarter, and the collection times are more than 12 times.
3. The method of claim 1, wherein the polyurethane functional material member is a polyurethane film.
3. The method of claim 1, wherein the polyurethane functional material member is a polyurethane film. The specific method for processing the spectral test images, the temperature test images and the brightness test images is as follows: Use a classic anomaly detector to process the spectral test images to extract spectral feature data; Use a classic edge detection operator to process the temperature test images and the brightness test images to extract temperature feature data and brightness feature data; Fuse the spectral feature data, temperature feature data and brightness feature data to construct a time series.
4. The method of claim 3, wherein the polyurethane functional material member is a polyurethane film. 5 Label the test shooting time for the time series to construct training samples.
Citation Information
Patent Citations
Method and device for measuring chemical oxygen demand in water by using fluorescence spectrometry and storage medium
CN115236045A