Research On Energy Consumption Management System Based On Non-intrusive Load Decomposition

With the rapid development of China’s modernization,almost all industries are inseparable from the use of electricity,and there are various types of electrical appliances.However,China’s energy consumption management system is not complete,and there is no experience in the management of electricity consumption.As a result,the annual power loss is huge.Therefore,in line with the principle of sustainable development,we plan to build a building power management system,The system can carry out unified management,scheduling,control,state analysis and data analysis of the electrical appliances connected to the system,so that the managers can more directly understand the energy consumption of the electrical appliances,and then give a more reasonable management scheme,so as to achieve the purpose of saving electricity.This paper mainly through the research of non-invasive load decomposition method,hardware design and software design,completed the development of energy consumption management system for non-invasive load decomposition.1.The load identification method of optimal state extraction and optimal neural network is studied.The original feature data is extracted by non negative matrix decomposition algorithm,then the peak point of maximum density is extracted by the density peak clustering algorithm.The maximum density peak point is used as the input of fuzzy C clustering,and the operation status points of each equipment are finally obtained,and the feature library template is established,The parameters of the two-way long and short-term memory network are optimized by using the seagull optimization algorithm.Finally,the results show that the method has a good improvement on the recognition effect,and the mean square error is also greatly reduced.2.Completed the design of the hardware part of the energy management system.Firstly,the performance index of the chip is determined to meet the functional requirements.At the same time,through the comparison of chip size,price and other factors,the selection of MCU,communication module,data acquisition module and other devices is finally completed.Finally,the chip design of STM32F103 and the design of data acquisition,program download,indication circuit,control circuit and other circuits are completed.3.Completed the software design.Firstly,the overall structure of the software system is introduced,its advantages and disadvantages are analyzed,and the specific scheme is explained,and the requirements and management scheme of the system are analyzed.The communication protocol of data link layer is selected and designed,and its different modes are compared,and the verification principle of the selected mode is explained.This paper has completed the construction of the prototype of the energy management system,completed the debugging of the prototype,and tested the performance and stability of the prototype.According to the test results,the system can collect and transmit information to the server in real time,and can monitor and control various devices in the management system in real time,achieving the expected results.