| Load monitoring is the basis of obtaining user’s detailed electricity consumption, analyzing user consumption behavior and spread of the work in energy-saving. Power supply enterprises to develop a reasonable demand response strategy through the analysis of users’ behavior, so that to guide users to use electricity in right way, ensuring the stability and economy of power supply. On the other hand, by learning the power supply situation, power grid policy and its own accurate electricity information, reasonable arrangements for users’ electricity consumption can leads to less energy consumption as well as expenses. Non-intrusive load monitoring method, which only needs to install the information acquisition device at the entrance of the monitoring system, analyze the data with appropriate algorithm when identifying and refining the power system, has become a developing trend since its great efficiency on monitoring and management.In this paper, a method of on-line parameter identification for household electrical appliances under the condition of non-invasion is studied, containing the load characteristics, performance analysis, load clustering and so on. An invisible platform is also constructed. Details shown as following.(1)Described the research background and significance of on-line parameter identification method for household appliances. Via researching the current development of load monitoring systems and load performance analyzing, combined with the user demand, the advantages & physical structure & working theory are concluded.(2)Established the load collection system which achieves the data collection of common electric equipment, analyzing the electric consumption performance with current, voltage, active power, reactive power, harmonic content, phase angle. With these prior training samples of multi-dimension characteristic, the unique feature of different loads can be identified.(3)Researched the transient state identification algorithm based on load switching state recognition algorithm to obtain the start point of steady process.Using principal component analysis method to reduce load characteristics dimensions of 8 typical electrical appliances, getting the optimized identify feature to construct the second cluster machine, first cluster machine through Fisher supervised linear discriminant design on Matlab.(4) Complete the visual platform of the online parameter identification system for home appliances by using the virtual instrument, achieved sign in, monitoring, situation identify, visible load monitoring, history data query and communication, showing the algorithm in this paper directly and convenient. |
PDF Full Text Request