Research On Real-time Online Monitoring Technology Of Electric Loads Based On IoT

With the rapid development of technology,people have put forward new requirements for the power system.Traditional power consumption has the characteristics of large loss,low safety and difficulty in troubleshooting.The resulting problems are getting worse.In the context of Io T(Internet of Things)technology and artificial intelligence,an intelligent electrical load monitoring system needs to be established in order to achieve scientific application of energy.The traditional electric load monitoring system focuses on the power information at the power inlet,but it ignores the correlation information between the power information and the type of electric load.Moreover,power information is often in the hands of power companies.The electricity consumption information of the electric load in the site,however,is still scarce for individual users.This paper aims to establish an associative management system for power inlet location,electrical load type,real-time power load and remote control to provide an Io T-based real-time online electrical load monitoring system for power companies and individual users.The main research work of this paper is as follows.(1)Overall design.The electric load monitoring system based on the Internet of Things consists of three parts: non-intrusive hardware node,cloud server and client software.The non-intrusive hardware node connects to the cloud server via Wi-Fi and transmits power information as well as control commands through TCP/IP protocol.The cloud server is responsible for data processing and instruction forwarding,including the execution of an electrical load identification algorithm,forwarding power data and electrical load operating status to client,forwarding switch control commands and power thresholds to non-intrusive hardware nodes.The client software provides a user interface to facilitate monitoring and control of electrical load from power inlets.(2)Hardware node design.The hardware circuit design of the system includes power management circuit(isolated AC-DC conversion circuit,non-isolated AC-DC conversion circuit and DC-DC conversion circuit),power metering circuit,optocoupler isolation circuit,relay circuit and Wi-Fi circuit.The hardware node is installed at the power inlet to convert the AC power into DC power for the node circuit.The power metering circuit is responsible for sampling the working current of an electric load,calculating the electric power and uploading it to the cloud server through the Wi-Fi circuit.The frequency of the pulse signal needs to be measured during power metering.In this paper,the “phase shift method” is designed to eliminate the ±1 count value error of the frequency measurement counter without additional time cost.This method can also increase the frequency refresh rate(3)The electric load identification technology based on transient and steady characteristics is studied.Then an electric load identification method based on PCA and KNN is designed.The method has the characteristics of small calculation amount and high accuracy.(4)Software design.This part includes the software design of the microcontroller on the hardware node,the software design of the cloud server and the software design of the client.(5)Implementation and testing of electric load monitoring system.This part completes the frequency measurement accuracy and power statistics experiment as well as the load identification accuracy experiment,during which data collection and training are performed on eight commonly-used household electric loads,and the recognition results and accuracy are given.The effectiveness of the system is finally verified.