This study focuses on the “blockchain + education” perspective, focusing on the integration of edge computing in the higher education resource sharing system. Through the benign interaction between blockchain and edge computing in the system data management system, the security and efficiency of data storage and transmission of shared resources in the system can be improved. In order to improve the performance of the system’s educational resource sharing, this paper utilizes the node identification model on the basis of the traditional PBFT consensus algorithm for the selection of master nodes and the monitoring of malicious nodes. Meanwhile, in order to ensure the balanced allocation of educational resources within the sharing system as much as possible, this paper utilizes the differential evolution (DE) algorithm for the balanced allocation of system resources and the educational resources within the system. The results of experiments and system tests show that the improved PBFT consensus algorithm (NR-PBFT) in this paper shows obvious superiority in tests such as throughput and latency. Although the educational resource allocation model performs poorly in the allocation of resources with larger technology such as digital books, the results for the allocation of teacher resources can effectively prove the effectiveness of the resource allocation model in this paper. In addition, the system test results also show that the system in this paper has good performance, and the introduction of edge computing can significantly reduce the packet loss rate of resource sharing, which has considerable application value.

Mental health issues have become a global concern. Aiming at the complexity of individual facial emotion expression in the task of analyzing mental health status, this study proposes a face emotion recognition method oriented to psychological intervention. The method integrates image recognition and sentiment analysis techniques, adopts Adaboost algorithm for face detection, generates an emotion region suggestion network based on face image recognition, and constructs an image sentiment classification network through feature map mapping and shared convolution. The method is then applied to the mental health recognition system. The model in this paper avoids the effects of individual and illumination differences. It has good face emotion recognition on several datasets, and the prediction accuracies are above 90%, especially for Happy emotion. In the comparison with other recognition methods, the recognition accuracy of this paper’s model is improved by 12.92% to 22.95%. The experiments show that the proposed face emotion recognition method can effectively predict the emotion of facial expression data in the mental health recognition system, and promote the assessment of individual mental health status and emotion management.

In order to optimize the design effect of cultural and creative products with non-heritage patterns, this paper uses image reconstruction algorithm and image recognition algorithm to process non-heritage problem patterns. By combining the processed non-heritage cultural patterns with consumer demand for cultural and creative products, non-heritage cultural pattern cultural and creative products are designed to meet market demand. On the basis of recursive network, we add multi-scale feature extraction module and attention feature fusion module, choose L1 loss function to optimize the details of image reconstruction, and construct image super-resolution reconstruction algorithm based on multi-scale recursive attention feature fusion network. And the image feature extraction network containing MSA module is designed, which is the fine-grained image recognition network based on multi-scale attention. The non-heritage cultural pattern dataset is established, and in order to optimize the recognition rate of non-heritage patterns, the image reconstruction based on multi-scale recursive attention feature fusion network is carried out on the non-heritage cultural pattern data. In view of the creative design strategy of non-heritage culture, the evaluation indexes of non-heritage cultural and creative product design are obtained from the consumer research, and the implementation suggestions of non-heritage pattern cultural and creative product design are derived based on the ranking of the importance of the evaluation indexes. The multi-scale recursive attention feature fusion network proposed in this paper achieves 34.89dB and 90.52% indicator scores on the Set5 dataset. For the design of cultural and creative products with non-heritage patterns, consumers make more suggestions in terms of functional differentiation, having a response rate of 21.58%.

The article calculates the average image entropy of the image domain, quantitatively analyzes the information has richness asymmetry in the task of digitally generating ink paintings, and constructs an asymmetric cyclic coherent ink painting digital generation model based on graphical algorithms. The model integrates a generative adversarial network, and the generator is centered on the Dense Block and replaces the residual block with a dense block to improve the characterization ability. The position fusion attention network is utilized to capture the main body region of the ink painting and combined with the edge extraction technique to extract the significant main body edges of the image and simulate the salient features of the ink painting strokes. The model is integrated into the teaching of “Children’s Ink Painting” course in a high school teacher, and students are instructed to use the algorithm to generate digital ink paintings to further explore the effectiveness of the teaching method. In this paper, the model is iterated for 30 times, and the total objective function converges to the minimum value of 0.85, and the measured values on PSNR, UIQM and UCIQE are improved by 4.44, 0.3 and 0.68 respectively compared with the optimal values of the comparison model, and the model can obtain the highest evaluation score (8) of the generated image at the fastest convergence speed (50 epochs), and the degree of overlap with the real image on the LPIPS distance is higher. After the experiment, the dimensions of digital pedagogical literacy level of the experimental class increased by 3.37 to 7.63 points compared with the control class and showed significant differences. As for the satisfaction of learning experience, students’ satisfaction with digital teaching resources is the highest, which is 4.70 points. The experimental results show that the model constructed in this paper has good performance of ink painting image generation and can be used as a digital teaching method for children’s ink painting course in high school teachers.

In this paper, the strain law of natural gas pressure vessel steel fatigue is firstly analyzed through the stress-strain curve and steel fatigue life curve, and the finite element model of natural gas pressure vessel is constructed by combining ABAQUS simulation software, and the fatigue performance of natural gas pressure vessel steel is analyzed from the cyclic softening behavior of the material and SN curve. Then the stress intensity factor theory combined with Paris formula is introduced to calculate the crack expansion rate, and ABAQUS and FTANC3D are jointly simulated to study the crack expansion law. Finally, the fatigue yield strength of the natural gas pressure vessel was analyzed based on the elastic-viscoplastic constitutive model combined with the finite element model. It is found that the stress intensity factor along the path under hydrostatic loading is larger than that under stresscontaining loading, but the difference in stress intensity factor is only about 1.42%. When the cracks of the natural gas pressure vessel extended to the vicinity of 20 mm, its crack extension rate showed a sharp downward trend. When the temperature comes to 900°C, the yield strength value of the steel of natural gas pressure vessel is only 280.42 MPa.Exploring the steel fatigue performance and crack extension rate of natural gas pressure vessel can help to better ensure the stable and safe operation of natural gas pressure vessel.

The article explores the method of diversified modeling of college sports track and field data, aiming to provide a basis for scientific training of college sports track and field. In this article, the diversified modeling of college sports track and field data is carried out by using multiple linear regression model, testing method and mathematical statistics method in order to analyze the sports characteristics and training needs of college track and field athletes. Using multiple linear regression model to analyze the influencing factors of track and field special movement patterns, then, on the basis of clarifying the training needs of track and field special movement patterns, combining the theoretical study of functional movement screening with the actual practice of track and field sports, carrying out the FMS test of the research object, and proposing the optimization plan of college sports track and field training after analyzing the results of functional movement screening of different track and field events. By using the multi-dimensional modeling method of college sports track and field data proposed in this paper to analyze the influencing factors of athletes’ track and field special action patterns, it is found that there is a significant medium correlation between the “torso forward swing and hip and knee rotation speed” in the buffer action link and the “torso extension speed” in the kick and stretch action link and the in-situ jump height. At the same time, there was a significant correlation between the common factor “trunk forward swing and hip and knee rotation speed” and the “trunk extension speed” in the push and extension link.

Using digital back camera to complete the traditional national costume image acquisition work, and then with the help of VOLO model to segment and colorize the image, the traditional national costume elements were successfully extracted. By fusing them with smart wearable devices, a detailed fusion implementation scheme is developed, which contains constraints and objective functions. In the context of numerical computation optimization, the fruit ϐly algorithm (FAO) is used to explore the fusion design scheme of the two in depth. The values of the four objective factors of the fusion design are 0.233, 0.232, 0.348, 0.144, and the ϐinal value of the objective function is 0.957, which indicates that the results of this paper not only can improve the comfort of the device and the user’s experience, but also can provide a new idea and method for the fusion of the apparel industry and the wearable device industry.

With the development of big data and education informatization, education reform and talent cultivation mode are facing digital reform. In this paper, the important feature selection algorithm based on random forest is used to select the relevant features that affect the application effect of teachers’ practice teaching cultivation and innovation mechanism, which lays the foundation for constructing the practice teaching data mining model based on Light GBM. Then the data processed by feature selection is preprocessed and standardized, and then the processed data is partitioned and the model is trained in turn to get the prediction results. The Light GBM-based practical teaching data mining model was compared with other classification models in different datasets, and the experimental results showed that the model in this paper has an advantage over other classification models in a number of evaluation indexes, with the highest accuracy rate of 13.07%, and the model data mining results accurately locate the open innovation experimental indexes that have a lower score of importance to students’ development, and provide a good basis for the optimization of teaching paths and students’ development. , which provides ideas for the optimization of teaching paths and the improvement of the impact of students’ future development.

Sparse decomposition has been generally emphasized in signal processing theory. In this paper, a nonelectrical signal feature dataset of key components of high-voltage DC converter valve is established by using principal component analysis to streamline the data volume. The compression-aware feature extraction algorithm based on polynomial matrix sparse coding is used to extract and collect the nonelectrical signal parametric data. Through the performance over the experimental signal analysis, it can be known that the eigenvalues of a total of 10 parameters, including the infrared temperature measurement results, the appearance, the presence of corrosion or dirt, and the presence of abnormal vibration and sound, are all greater than 1. Therefore, these 10 parameters are identified as the key parameters. When the number of measurement points is between 64 and 200, the algorithm in this paper can satisfy the need of feature extraction when the signal length is insufficient, compared with the traditional approach. In the empirical analysis of the vibration signal as an example, the method of this paper can effectively extract the frequency and time domain of the vibration signal.

As the core equipment of high-voltage direct current transmission system, the operation status of the converter valve directly affects the safety of the power grid. In this paper, we first construct a multisource data fusion system to realize the error-free fusion of fault information parameters. Then, combined with the random forest algorithm, the time-varying law of the electrical characteristics of the converter valve based on harmonic theory is extracted. Finally, the collected time-varying laws of electrical characteristics are input into the constructed Random Forest particle swarm optimization model, and the trained model is used to monitor the status of the converter valve. In the simulation experiment, the ±800kV UHV DC transmission system is built by PSCAD/EMTDC software, from which the current waveforms are collected when the converter valve fails, the time domain features of the current are extracted, and the obtained converter feature indicators are selected using the Random Forest algorithm, and 10 important features will be finally identified to construct the converter valve feature indicator set, and input into the Random Forest Particle Swarm Optimization model and the other comparative models for training and testing. The accuracy of this model is 97.5%, which is better than other comparative models. The study provides a high-precision solution for converter valve condition monitoring and effectively extends the application of multi-source data fusion in power equipment.