In this paper, first of all, the data preprocessing of ethnic patterns is carried out through image segmentation and grayscaling processing methods, and then the image processing technology is applied to the feature extraction of ethnic dress patterns, and the improved SIFT algorithm is used for the feature extraction of images. The original DCGAN algorithm feature extraction ability is weak generates style picture fuzzy, the effect of the problem of poor, proposed the use of 32-layer deep neural network with residual structure instead of the original 5-layer shallow feature extraction network, significantly enhanced the algorithm’s feature extraction ability, enhance the model of the style migration effect. By introducing the objective evaluation index PA of the improved SIFT algorithm, the algorithm was compared with other algorithms, and the segmentation algorithm experiments were carried out with the local patterns of several images, and the results of pixel accuracy PA were obtained to be greater than 0.95, which confirmed that the improved SIFT algorithm was able to realize the accurate extraction of the contours of local patterns. In terms of pattern quality evaluation dimension, the subjective average scores of the amateur group and the expert group are 4.87 and 4.89 respectively, indicating that the ethnic patterns generated by the algorithm of this paper have reached a high standard in quality.

With the accelerated pace of society and the increasing pressure of competition, the issue of mental health has received increasing attention. Especially in the field of education, students’ mental health status directly affects their student outcomes and overall development. The aim of this study is to design a mental health status monitoring system based on large-scale data streaming computation, to realize dynamic real-time monitoring of individual mental health through multi-source data acquisition and efficient algorithm processing, and to explore its application in educational scenarios. Sliding window algorithm and Hidden Markov Model are used to analyze and process the collected multi-source data such as physiological signals, and the experimental results show that the system is able to significantly test the difference between people with high and low scores on psychological test scales in the monitoring of mental health status, and it can provide educators with valuable decision-making support and help students’ mental health education and intervention.

Solving the health problems of key populations such as people with disabilities is an important way to realize universal health coverage and promote social equity. Sports as the main means of rehabilitation for disabled people at present, this study clarifies the concept of disabled people and sports rehabilitation, and uses empirical investigation to analyze the plight of cruel sports, laying a realistic and theoretical foundation for this paper based on high-dimensional rehabilitation data optimization. In the process of sports rehabilitation exercise for the disabled, it is generated through the modular control of multi-skeletal muscles by the central nervous system as well as the regulation of neural oscillations. And the surface EMG signal is the combined effect of superficial muscle EMG and electrical activity on the nerve trunk on the skin surface, which belongs to the high-dimensional data characteristics. For this reason, this paper constructs a rehabilitation exercise muscle synergy model using matrix decomposition for the rehabilitation of disabled people’s sports. The data were refined in the time domain by adding time windows, and then the data were decomposed into targets based on the non-negative matrix decomposition method to extract the muscle synergy features in each time window, so as to analyze the muscle synergy differences in different exercise processes and different feature frequency bands, and to further obtain the muscle synergy law during exercise and the physiological change mechanism of the nervous system during exercise control. Finally, the experiments were carried out in both non-electrical stimulation and electrical stimulation modes, and the results showed that the number of muscle synergism in wrist flexion and extension was the same in both modes, which was 3. However, the number of synergistic pairs of muscles in the electrical stimulation mode was significantly increased. It also proves the effectiveness of the method of this paper on the analysis of muscle synergy of multi-channel surface EMG signals, which provides a new method for exploring the muscle synergy characteristics and the control mechanism of rehabilitative movement in the process of disabled people’s sports.

In order to cope with the damage of urban electricity and the dilemma of residents’ electricity consumption caused by flooding disaster, we study the dynamic planning of intelligent operation and maintenance equipment scheduling and distribution network restoration under flooding disaster. Consideration is given to both pre-disaster deployment and post-disaster scheduling levels, while dynamic planning is carried out for collaborative repair and energy storage scheduling to construct a scheduling model with multi-source collaboration. Based on this, a multi-resource cooperative post-disaster recovery strategy for distribution networks is further proposed. The usability of this paper’s multi-source cooperative strategy is studied in depth through case analysis. Among the six Cases of the simulation experiment, the total cost in Case 1, which is operated and restored according to the strategy proposed in this paper, is the lowest, which is only 257080.2 RMB. The maximum, minimum, and average values of the solution time of the multi-source cooperative strategy are much faster than those of the comparison methods, and it has obvious advantages in fast decision making. The multi-source synergy model in this paper is able to recover all the loads within 285 min, while the finite synergy model takes 330 min. The multi-source synergy model was able to recover 7,500 kW of load, while the limited synergy model was only able to recover 6,850 kW. The multi-source cooperative model has strong applicability.

Organic light-emitting diodes (OLEDs), as a new generation display and lighting technology, are critical for charge transport and luminescence efficiency enhancement. The study determines the potential, electron and hole concentrations in OLED devices based on Poisson’s equation and current continuity equation, and fits the charge transport process in the devices with the drift-diffusion model. The differential equations are solved by improved Euler’s method and iterative solution method to simulate the operating state of the OLED device. In conjunction with experiments, the enhancement effects of the OLED devices optimized based on the differential equation model in terms of charge transport and luminescence efficiency are analyzed. The optimized device and the comparison device exhibit the same partial pressure and a largely overlapping luminescence curve at 450~460 nm, but the optimized device lifetime and brightness are better than the comparison device. The charge transfer efficiency of the optimized device exceeded 99.99%, while that of the comparison device was lower than 95%. In addition, the light extraction efficiency of the optimized device is more than 20% higher than that of the comparison device, and it has the highest current efficiency, i.e., the optimized device has a better luminescence efficiency. The differential equation model is used in OLED devices to describe the processes of charge transport, optical properties, etc. The model can be used to systematically optimize the material properties and improve the overall efficiency of OLED devices.

This paper proposes to design the power meter based on TMR current sensor, screen the chips that meet the requirements of the power meter, and stipulate the technical specifications and technical parameters of the power meter based on TMR current sensor. Design the system structure of power meter with TMR current sensor including MCU module, storage module, communication module and so on. And design the main and vice system clocks in the single-phase energy meter with TMR current sensor. Analyze the design of signal acquisition module, bias adjustment and temperature compensation module, communication module and circuit protection module in the current monitoring system. According to the characteristics of the TMR sensor, establish the objective function, improve the GWO algorithm, and optimize the design of the multi-stage magnetic ring structure current sensor. The performance parameters of the TMR sensor are analyzed, and the DC current test and AC current test are conducted to verify the performance of the TMR current sensor measurement module. The accuracy, precision and linearity of the current measurement module are tested, and the relative error between the actual current value and the theoretical current value derived from the formulae in the DC current test and the AC current test are controlled within 5% in the TMR current measurement system. The measurement system based on TMR current sensor meets the current measurement requirements.

Artificial intelligence technology can effectively improve the quality and efficiency of industrial design and manufacturing, so the study takes Shuangdun Carved Symbols of cultural products as an example, utilizes the generative adversarial neural network to carry out style migration processing in the design of Shuangdun Carved Symbols and their products, and constructs the DCGAN model to assist the design and generation of Shuangdun Carved Symbols of cultural products. After semantic analysis of the color symbols of Shuangdun Carved Symbols products generated with the aid of DCGAN model in this paper, quantitative and qualitative measurements are carried out. Users of Shuangdun Carved Symbols products rated the products after the style migration significantly higher than before the migration in terms of volumetricity, distance, emotion, character, and texture.CycleGAN and DCGAN models achieved the best overall results in terms of PSNR, SSIM, FID, and KID indicators. The DCGAN model with added spectral normalization and Res2Net outperformed the CycleGAN model in the ablation experiments. The overall user rating of the Shuangdun Carved Symbols product designed by the DCGAN model in this paper is 4.24, and the product has obtained more satisfactory evaluation results.

Under the new situation of continuous and stable development of China’s economy, large products have extremely high requirements on transportation safety due to their high price, complex transportation technical requirements, which determines that large products should be delivered to customers in the safest and most economical way, which poses a difficult problem for decision makers to choose the optimal path. In this paper, we constructed an intelligent approval framework for bulky transportation, made technical and economic analysis of transportation routes, and established a multi-objective optimization mathematical model for path selection of bulky transportation vehicles. A hybrid genetic algorithm incorporating greedy strategy is proposed to solve the problem, which strengthens the ability of the algorithm to jump out of the local extremes and selects the optimal chromosome in the final population as the resulting optimal solution. The results of the approval and optimal route planning for bulky transportation are verified by the method of example experimental analysis. The volume of bulky transportation increases with the increase of years until 2023, and the GDP, value added of tertiary industry, total population, and road mileage are 1015987.54, 553948.15, 140563, and 536.48, respectively. In the instances where the number of orders is 2000 or more, the transportation distance, the maximum number of service bundles of orders on the route, and the maximum service hours of vehicles the mean values are 50, 3.56, and 14.33, respectively. According to the constructed mathematical model, the optimal line for the bulky transportation scheme is 0→2→4→7→8, and the total transportation cost is 670,500,000 yuan, of which the transportation costs are 116,500,000 yuan, 320,000 yuan, 151,000,000 yuan, and 83,000,000 yuan, respectively.

Aiming at the problems of unfixed switching frequency and complicated calculation in the control of permanent magnet synchronous motor, a permanent magnet switch FNN-PID control strategy based on deep learning technology is proposed. Based on the vector control of permanent magnet synchronous motor, the resonant pole inverter is combined with permanent magnet switch control, and then the fuzzy neural network and incremental PID algorithm are used to construct the optimization strategy of permanent magnet synchronous motor switching frequency FNN-PID control. And combined with the finite element simulation software, the permanent magnet switch finite element model is constructed, and the effectiveness of the FNN-PID control strategy is illustrated by verifying the permanent magnet switch control strategy and the temperature rise curve change. When using the FNN-PID control strategy, the electromagnetic torque quickly reaches stabilization near the given torque of 9 N-m after 0.03 s of startup, and the permanent magnet switch frequency of the FNN-PID control strategy is reduced by 24.04%. The difference between the measured maximum winding temperature and the calculated maximum temperature under rated operating conditions is less than 9°C, and the permanent magnet switching loss is reduced by about 35% with the FNN-PID control strategy compared with the traditional MTPA control strategy. Therefore, the combination of deep learning technology and finite element analysis can explore the optimization effect of PM switches from the strategy and application dimensions and provide research ideas for the stable operation of PM switches.

The ring network cabinet of the distribution network is an important part of the urban power system, and its operation state directly affects the stability and reliability of the power system. In this paper, a deep learning algorithm is used to analyze and process the partial discharge signal, and a permanent magnet fast ring main unit partial discharge detection and fault identification model based on improved DBN-LSTM is proposed. By analyzing a large amount of local discharge signal data under normal operation and fault conditions of ring main cabinet, and using these data to train a deep learning-based fault prediction model. The performance of the improved DBN-LSTM model is tested by combining the defect spectrograms of four typical ring network cabinet partial discharge models and compared with other algorithms. The proposed model has good effect on fault identification of ring network cabinet, with a combined identification accuracy of 98.41%, and the overall identification performance is better than both BP neural networks and SVM classifiers. The prediction accuracy of the fault prediction model also reaches 88.52%, and the experimental results of the method in this paper are more satisfactory.