A non-uniform clustering model based on energy harvesting is proposed for the “hot zone” problem in oil and gas IoT( Internet of Things). The model takes into account the residual energy of nodes, the distance between nodes and base stations, the density of neighboring nodes and the distance between nodes and RF (Radio Frequency) signal sources when setting the node competition radius. It improves the non-uniform cluster formation effect, elects double cluster heads to balance the energy consumption of cluster heads, uses multi-hop routing in the data transmission stage to balance the energy consumption of the network, and introduces wireless energy harvesting technology to extend the network lifetime. The simulation results show that this algorithm can more effectively balance the network energy consumption and extend the network lifetime compared to the traditional algorithm.

In order to optimize the CCHP(Combined Cooling Heating and Power) supply system reasonably, a system of efficient installations of saving is established. It improves system efficiency, and realizes the cascade utilization of energy. On the basis of existing research, a waste heat recovery system is built. It contains a generator set and refrigeration unit. According to the partial load characteristic of the system on the unit in the system modeling alone, the original particle swarm optimization algorithm is optimized, and an example is introduced to verify it. The results show that the system has a good effect on the optimization of the algorithm, and the improved algorithm ensures the stability and practicability of the system operation.

In order to improve the accuracy of load forecasting results, a short-term load forecasting model based on improved fractal theory is designed. The date similar to the meteorological data is selected as the reference date, and the reference date is analyzed by re-scale-range method to determine that the reference date had fractal characteristics. The Iterative compression factor is calculated according to the fractal interpolation interval and the IFS (Iterative Function System) is established to establish the fractal interpolation Function of the reference date. The moving average function is used to process the data, and OLS (Ordinary Least Square) is used to establish the data fitting equation, and the time data is put into the fitting equation to calculate the predicted data. After simulation comparison experiment, the MAPE ( Mean Absolute Percentage Error ) of load data predicted by the improved prediction model decreases by 0. 26 compared with the previous prediction model. It is proved that the short-term power load forecasting model based on improved fractal theory can effectively improve the accuracy of load forecasting results. 

Aiming at the problems of large number of parameters, large size and poor anti-noise performance in traditional fault diagnosis models, a bearing fault diagnosis method based on MSDS-CNN( Multi-Scale Depth Separable Convolutional Neural Network) is proposed. The input signals are processed in parallel by using depth separable convolution of different scales to obtain multi-scale information and ensure the lightness of the model. The Dropout layer is introduced to improve the anti-jamming ability of the model, and the global average pooling layer is used to replace the full connection layer to reduce the number of model parameters. The experimental results show that the diagnostic accuracy of this method is up to 99. 6% , which is higher than other methods. The model has fewer parameters, smaller size and lighter weight. It also has good diagnostic accuracy under noise interference. 

The operation of power system is influenced by many factors and the power load data present strong volatility and instability affecting the accuracy of short-term load forecasting of the power grid. Traditional power load forecasting methods have increased prediction error when affected by nonlinear factors such as policy, weather and holidays, so a combined model strategy is proposed. Firstly, the original data is decomposed into several components by EEMD (Ensemble Empirical Mode Decomposition), and the components are divided into two groups according to the different information content of each component data, and the prediction is carried out by using back propagation neural network and short-and-long memory network respectively. A SSA ( Salp Swarm Algorithm) is used to optimize the number of neurons and the lag term of input variables in each component prediction network, and the final combined forecast model of EEMD-SSA is obtained. Finally, a real in-situ data is used with this combined model to forecast power load. The experimental results show that this combined model has better prediction effect than single network model and other models. 

Aiming at the problems of difficult manual extraction of motor rolling bearing signal features and poor fault classification effect, an improved GoogLeNet convolutional neural network is proposed by combining the traditional GoogLeNet model unit and the dense connection idea. The proposed model is applied to the fault diagnosis test of motor rolling bearings. After grouping and labeling the original data, it is directly input into the improved model for training, and finally the test set is input into the trained model to test its classification accuracy rate. The entire diagnosis process does not require manual feature extraction, which avoids the difficulties and errors caused by manual fault extraction, greatly simplifing the fault identification process, and proving the feasibility of the improved GoogLeNet model in fault diagnosis. Finally, the proposed model is compared with the traditional GoogLeNet model and other typical models. The comparison results show that the improved GoogLeNet convolutional neural network model has the characteristics of higher accuracy, strong feature extraction ability, fast convergence speed, and stable performance than the traditional GoogLeNet model and other comparison models. 

In order to solve the problem of multi-peak output power of photovoltaic power generation generated by local shadows, a compound algorithm combining gray wolf optimization algorithm and disturbance observation method is proposed. Levy flight module and greedy strategy are embedded into the grey wolf algorithm to achieve optimization. The composite algorithm makes use of the accuracy of gray wolf optimization algorithm and the rapidity of disturbance observation method to accurately track the maximum power point of photovoltaic power generation under multiple peak values. The simulation results are verified by Simulink experimental platform. The Simulation results show that the proposed algorithm has better tracking performance than the traditional gray Wolf algorithm in accuracy and rapidity. 

In view of the problem that the amount of data received and processed by the three-dimensional lightning detection system is several times higher than that of the two-dimensional lightning detection system, and the warehousing process of positioning, return stroke and status data is much more complex than that of the two- dimensional system, a three-dimensional lightning detection data processing and storage system is developed using Oracle database and PL( Procedural Language) / SQL( Structured Quevy Language) developer integrated environment. The basic information of the station is stored by designing the location information table of the three-dimensional lightning locator. The lightning location result information table is designed to store the processed positioning data and return stroke data. The latest state information table is designed to store the latest status information of the locator. And this information can be used to determine whether the device is in normal state. Through the construction of an experimental platform of Hubei three-dimensional lightning detection network, the function and operation state of the system are verified. The results show that the design and implementation of the processing storage system meet the mission requirements of the data processing and storage function of the three-dimensional lightning detection system.

In order to solve the problem that the accuracy of multi-segment short-term load forecasting is low due to the strong chaotic characteristics of load data and the influence of noise in the power system, a power system multi-segment short-term load forecasting method based on improved particle swarms is proposed. Based on the historical data of the power system, the concept of particle aggregation is introduced, the solution space is established, the global load data is searched in the space, and the original data is introduced into the solution space to determine the data distribution range. The optimal objective function is established, the adaptive load particle weight is calculated using the law of linear decline, and the particle weight is continuously approached to the optimal value with the iterative update function. The local prediction function and the global prediction function are integrated with the improved particle swarm prediction law, and the matrix is adjusted with the largest decision weight coefficient to complete the load forecast. Simulation experiments prove that the proposed method has strong ability to judge and analyze load data, good adaptability, and high degree of fit between the predicted results and actual values.

Because most of the current research on distribution network reconfiguration is aimed at small and medium-sized networks, and there is less research on large networks, an optimized GA( Genetic Algorithm) algorithm is proposed. In the large-scale distribution network reconfiguration, the gene operation process is optimized, and the elite strategy is introduced to retain the genetic information of the parent optimal individual to the greatest extent. In the process of variation, the “ local variation冶 is introduced again, and the offspring population is more diverse. The correlation coefficient is introduced into the reconstructed objective function to simplify the computational complexity. The algorithm is applied to IEEE118 node system. Compared to the results obtained by MC(Monte Carlo) simulation method, the reconstructed network loss and voltage deviation are obviously better, which proves the feasibility and effectiveness of the algorithm in large-scale distribution network reconstruction.

The traditional deadbeat current predictive control depends on accurate discrete power grid mathematical model, which leads to steady-state error of the current, especially the system instability when the inductance changes beyond a certain range. There is sampling delay in the all digital control, and the system robustness is poor. An improved deadbeat predictive control method combined with repetitive control is proposed to effectively eliminate the command error and disturbance error of grid connected current and provide high-quality steady-state waveform. Simulation results show that the influence of sampling delay caused by traditional control methods is effectively suppressed. It has the advantages of fast response, high steady-state accuracy and small current distortion rate (THD: Total Harmonic Distortion), and avoids the disadvantages of overshoot and oscillation.

 In order to identify chronic, acute, brown dot and white dot, four types of rice blast diseases early and accurately, a deep neural network is constructed by combining sparse automatic encoder and SPSO-SVM (Switching Particle Swarm Optimization Support Vector Machine). Compared with other algorithms, the neural network needs to input a large number of images, the autoencoder can extract the most representative information in the original image, reduce the amount of information in the input, and then put the reduced information into the neural network to learn, greatly reducing the difficulty and time of learning. Firstly, the input data is encoded, decoded and reconstructed by sparse automatic encoder to learn the hierarchical features of rice blast leaf spots, and the sparse condition constraint is added to the automatic encoder to compress the hidden layer, so as to learn the higher-level hidden features. Secondly, the support vector machine optimized by switching particle swarm optimization is used to identify the types of rice blast. The open Kaggle rice disease image database and the actually collected rice blast image are used as the data set. 350 images of each type were selected to form samples, and each image is normalized to 4096 dimensional vector. 80% of the samples are randomly selected as the training set and the remaining 20% are used as the test set. Through 10 cross validation, the average recognition accuracy of the test set is 95. 7% . The experimental results show that the proposed method can effectively identify the early disease of rice leaf blast from the features of disease spots, which is of great significance for the early prevention of rice blast.

When quantum optimization algorithm is applied, the difference of optimization control methods will cause the convergence time of optimization algorithm to be too long. Therefore, a quantum optimization algorithm for decentralized performance of nonlinear chaotic network systems is proposed. The piecewise Logistic chaotic mapping method is used to obtain the quantum initialization position. Based on the fitness division of quantum populations, the corresponding evolutionary models are established for the top and bottom populations respectively. Using multiple nonlinear continuous time subsystems, the nonlinear chaos optimal control method is designed and the optimal control scheme is determined. Finally, the premature convergence judgment mechanism is introduced to obtain the calculation results of the optimal solution. Experimental results show that the convergence time of the proposed optimization algorithm is significantly reduced compared to the traditional method.

At present, the fixed time operation mode is adopted during the job scheduling of tunnel cleaning robot. This would waste the power and the fire resources and affect the driving safety. Aiming at this problem, an intelligent scheduling method of tunnel cleaning robot is presented. Firstly, the time varying operation character of tunnel infrastructure is described according to the influence factors, such as traffic flow, temperature, humidity and so on. Then, considering the operation performance of infrastructure, the operation cost of robot, the impact on tunnel traffic flow and so on, the cascading optimization scheduling model of tunnel cleaning robot is constructed. The scheduling cycle and operation time of cleaning robot are optimized respectively to realize the efficient operation of robot. In the experiment, the proposed method is compared with the commonly used fixed time scheduling method. The results show that at the same operation cost, the infrastructure operation performance increases 6% by using the proposed method compared to the fixed time scheduling method.

In order to solve the problem of low efficiency and poor stability in dispatching massive data in power system, a scheduling algorithm of massive power data based on grey fuzzy prediction is designed. Considering the real-time and reliability of power system task scheduling, the power task scheduling strategy is formulated. The deadline and value of scheduling task are selected as characteristic parameters, and the load rate is calculated to measure the actual load of service nodes to complete the load balance distribution of power system. The grey fuzzy prediction algorithm is used to schedule the power data, the future trend of a single existing storage block is predicted according to the progressive function, and then combined with the priority of the task to realize the automatic scheduling of massive power data. The experimental results show that the proposed algorithm can make the collaborative scheduling and resource allocation in a short time, ensure the stable state of the data platform, improve the data scheduling efficiency and enhance the scheduling stability. 

The number of image samples of rare fundus diseases is small, which is difficult to meet the needs of deep network training. A recognition model with MFPD(Multi-representation Fused with Prior Distributions) for rare fundus diseases is proposed. Based on the pre-training model, fine tune the embedded model to obtain the feature embedded prior distribution, map the embedded features to different spaces, and extract the image features from different perspectives. On the basis of cross entropy loss, divergence loss is added to increase the difference of different perspective features, make efficient use of rare disease image information, and reduce the impact of small sample size on the model. The experiment uses the OPHDIAT(Ophtalmologie-Diabˋete-Te′le′me′decine) fundus image dataset to compare this method with other methods. The experimental results show the effectiveness of this method.

In order to generate concepts efficiently, a concept generation algorithm IUICG(Incremental Updating Intent Based Concepts Generation) is proposed. First, the attributes in the formal context are set as task attributes one by one, and the concept search space is divided into privious concept set and newly-add concept set by the task attributes, which improves the search efficiency. Secondly, the concept operation rules are proposed, in which the extent filtering rules avoid the search of concept space by invalid extent, and the intent update rules and newly-add concept rules improve the concept generation speed. The experimental results show that the time performance of the IUICG algorithm is better than that of comparison algorithm on different types of data sets. The IUICG algorithm has nearly linear time complexity on data sets where the number of objects is much greater than the number of attributes. 

In the work of scientific and technological literature retrieval, it is very important to give a reasonable correlation ranking from a professional point of view. The traditional PageRank algorithm uses the method of evenly distributing similarity weights, but this method will cause the unreasonable results of literature ranking. Therefore, an algorithm combining deep learning method and PageRank is proposed to improve the reliability of literature relevance ranking. Firstly, the Siamese BERT ( Bidirectional Encoder Representation from Transformers) network with attention pooling is used to calculate the similarity between literature and citations, and then the similarity between literature and citations contained in literature is normalized. Finally, the normalized similarity is used as the distribution weight to calculate the ranking results of citation network. The experimental results show that compared with the traditional PageRank algorithm, the correlation of the retrieval results of this method is improved by more than 6% , which is more suitable for citation network analysis of scientific and technological literature.

 In order to better grasp the research hotspots and theme evolution of virtual reality, the application status of scientific papers on virtual reality funded by NSFC ( National Nature Science Foundation of China) projects in China was analyzed by using bibliometrics and itginsight. The results show that virtual reality research is in a period of rapid development, and it is expected to continue to increase during the 14th Five Year Plan period. School of architecture and urban planning, Tongji University, School of resources and civil engineering, Northeastern University, and State Key Laboratory of virtual reality technology and systems, Beijing University of Aeronautics and Astronautics have the greatest academic influence Journal of system simulation, modern electronic technology and Chinese Journal of image and graphics have the most papers. The research topics mainly involve virtual reality, visualization, key technologies, realism, digitization and other research hotspots. With the change of research topics, virtual reality has always been a hot topic, and new concerns such as visualization and robotics have emerged in the later stage. 

In order to further solve the problems such as lack of training data and the wrong makeup area in the existing facial makeup transfer methods, an automatic face makeup migration method based on the generation countermeasure network is proposed. This method constructs the objective function of generative adversarial networks, and achieves the generator by encoder-decoder neural network. Meanwhile, it constructs the discriminator based on multilayer convolutional neural network. The training algorithm adopts alternating optimization. The results of simulation experiment and method comparison show that this method keeps the facial structure, and reflects the reference makeup style as much as possible, achieves a more harmonious makeup effect, has better comparative advantages and visual effects, and provides a new idea for the automatic facial makeup transfer technology.

In early warning of the key nodes for power grid project, the unique outlier characteristics are considered. In order to solve the current problems of large risk warning errors, low accuracy and stability. An early warning algorithm of key nodes of power grid project based on outlier detection is proposed. Using the measurement index of outlier early warning, the task importance index of static outlier and outlier is calculated. Using analytic hierarchy process and entropy weight method, combined with multi index fusion weighting, the characteristics of key outlier nodes are extracted to complete the identification of key nodes. K-means is used to cluster the early warning process of key nodes of power grid. The fusion weight characteristics of key nodes of power grid are introduced into the outlier detection system to analyze the data output results, obtain the optimal clustering value, and realize the early warning of key nodes of power grid project. The experimental results show that the proposed method has high stability and accuracy, and can effectively reduce the risk of early warning error. 

Because the existing algorithms failed to cluster the power data, the recall rate decreases. Therefore, a refined verification algorithm for power multilevel redundant data in variable load mode is proposed. In the variable load mode, through the sparsity of power signal, the transformation coefficient is linearly projected to the low-dimensional observation vector to solve the sparsity optimization with high probability and complete the data acquisition. The genetic clustering optimization algorithm is used to cluster the collected power data. According to the clustering results, the analytic hierarchy process is introduced to obtain the correlation between different data, and the refined verification criterion of power multilevel redundant data is constructed, and the data verification is completed through the verification criterion. Experimental results show that the proposed algorithm can effectively reduce the verification time and extra storage overhead, and increase the recall ratio and the accuracy of verification results.

In order to accurately control financial risks and reduce risk losses, a prudent financial risk assessment method based on wavelet analysis and nonparametric estimation is proposed. A risk assessment model based on the inclusion probability distribution is established, and the financial risk coefficient through two inclusion probability distribution families is visually displayed. A parameter variable determination method based on wavelet analysis and nonparametric estimation is designed to clarify the trend of financial risk, and nonparametric estimation is used to obtain the probability distribution results of risk, so as to obtain the loss value of financial risk and provide important parameter basis for prudent assessment. The Monte Carlo algorithm is used to calculate the Monte Carlo integral of sampling points with stable distribution to improve the efficiency of prudent evaluation and capital utilization. The simulation results show that the model can effectively control the financial risk accurately, minimize the risk loss, and greatly enhance the prudent evaluation of financial risk.

The computation of Weil pairing of elliptic curves over finite fields is of great significance to the application of public-key cryptography. The research of Weil pairing focuses on theoretical research, but pays little attention to practical application, which leads to the need for new methods to support some theoretical research. For this reason, the Weil pairing calculation method is given, the point selection problem of Miller algorithm over finite field is pointed out by examples, and the difference between two different methods using Miller algorithm is analyzed. Through Miller algorithm, the limitation of MOV( Menezes-Okamoto-Vanstone) attack discrete logarithm is pointed out. The practical analysis shows that the Weil pairing of elliptic curves in finite fields are of small order and not very effective for large suborder.

In order to enable undergraduates in digital geological science to master the basic process of geological modeling, simulation and visualization, blender software, true terrain and other auxiliary plug-ins are used to carry out the teaching of model construction and dynamic simulation of geological evolution process for undergraduates in digital geological science, so that students can personally complete the whole process of geological modeling, and face the evolution of the earth for hundreds of millions of years. The actual results show that this mode can deepen the students’ mastery of the professional knowledge of digital geological science and more intuitively understand the formation law of geological phenomena, and play a very positive role in enriching the methods of geologic teaching and cultivating students’ professional quality. 

As the last link of express logistics, terminal pickup is also a link of direct contact between express logistics and customers. However, the current express terminal pickup mode, especially for areas with dense populations and large parcel traffic such as universities, still has problems such as low pickup efficiency and high labor costs. In response to this situation, we propose a new unmanned delivery terminal mode. This model uses STM32 single-chip microcomputer as the control system, and uses mechanical devices to automatically pick up parts under unmanned conditions. The terminal model machine is designed and manufactured. Experiments have shown that the express terminal mode improves the efficiency of picking up items, while reducing the labor intensity of the station staff, which meets the expected functions and requirements.