| Focusing on the development of electric power in China,from candles and kerosene lamps in the early days of liberation to the popularity of light bulbs in the 1970 s,to the popularity of household appliances such as TV,refrigerator,washing machine,etc.,the rapid development of the country has brought us high-quality material and cultural life,promoted the rapid development of electric power in China,but also brought more challenges and tests.Nowadays,with the continuous development of big data technology,Internet of things technology and intelligent technology,hundreds of millions of data are generated in the power industry every day,but they are not fully utilized;the power data in various regions are managed dispersedly,and the fault phenomenon of power grid data also causes the lack of relevance between the data.When the cross regional power grid dispatching is carried out,the staff cannot real-time In the process of dispatching,if the dispatching personnel want to obtain the required information,they need to do many times of searching and manual screening to find the original content,which requires high skills and knowledge of the workers.Therefore,to build a unified data access platform for the whole grid data is the key to improve the utilization rate of grid data and the efficiency of grid dispatching.Aiming at the goal of building the whole business data management platform for power grid dispatching,firstly,the work of data sorting is completed;secondly,the sorted data is integrated into the relational database,at the same time,the knowledge graph of power grid field is designed,and data transfer is realized to generate the visual knowledge map;finally,the project application is based on the search engine design based on the data knowledge graph of power grid field In consideration and implementation,the whole business data management platform of power grid dispatching can be built.In this paper,the prediction algorithm is improved to deal with the problem of electric power data prediction in the transmission data section in the process of data processing.By adding CNN convolution neural network layer to the LSTM time series prediction algorithm to extract data features,due to the sudden change of electric power in some special cases,a correction module is added.The final prediction result is the sum of the joint prediction result and the data value on the opposite side of the grid,and through experiments to verify the rationality of the algorithm,and reduce the prediction error.In view of the construction of knowledge graph,this paper selects neo4 j graph database as a transit database,extracts the equipment information from dameng database and transfers it to the designed node structure in neo4 j graph database,and completes the extraction and transfer of equipment relationship through business logic.Aiming at the phenomenon that equipment information such as plants and stations and the information between equipment are not easy to obtain,the grid information search engine is designed and implemented to improve the friendliness of the project platform to the staff.The B/S model architecture is adopted,the front-end search engine interface is written in Java,and the key words entered are matched with the node values in the background neo4 j database The attribute value of the matching node and the core content of the node whose relationship degree is within one degree are returned to the browser as the result of this search,and further returned to the user to complete this information search.The design and application of the search engine can greatly improve the work efficiency of the staff in the operation of obtaining data information,and also can complete the power grid scheduling more quickly.At the end of this paper,the completed work of the project is summarized,the future work of the project is prospected,and the next research direction of the project is further elaborated,It has done some promotion work for the follow-up work of the project. |
PDF Full Text Request