An Intelligent Algorithm for Non-Intrusive Appliance Load Monitoring System
Monitoring electrical power consumption has become an important research issue in order to reduce electricity expense and avoid unnecessary electrical operation. This study is based on the architecture of Non-intrusive load monitoring (NILM) system to mon
- PDF / 186,239 Bytes
- 5 Pages / 439.363 x 666.131 pts Page_size
- 57 Downloads / 224 Views
1
Introduction
In recent years, for the sake of saving energy and environment protection, people put a lot of efforts to avoid unnecessary electrical operation [Chang, Lin, Chen, and Lee, 2013; Anderson, Berges, Ocneanu, Benitez, and Moura, 2012]. A general house will set up a traditional power meter, which records only consumption power for the current house situation. Smart digital meter allows users to understand the situation of electricity utilization within the house. It provides power consumption information, such as active power, apparent power, voltage and current record by time. The information will be saved in a data base and analyzed by artificial intelligent algorithm to identify current operational appliances in the house [Chang, Lin, Chen, and Lee, 2013; Wang, Wang, Drummond, and Ahmet Sekercioglu, 2013]. NILM system requires an accurate algorithm to analyze the power information in order to accurately extract the electrical characteristics (also known as power signature), to achieve the effect of electrical identification. Hart [Zhenyu and Guilin, 2011] proposed an idea of a finite state machine that the electrical characteristics were extracted from the steady state load power consumption data. Although this method performs well in the sample appliance, it cannot monitor small appliances or appliances that are always on or have non-discrete changes in power [Zhenyu and Guilin, 2011; Chang, Chen, Tsai, and Lee, 2012]. There are many research papers use artificial neural network theory to improve recognition performance in NILM system. Chang uses the transient state of active James J. (Jong Hyuk) Park et al. (eds.), Future Information Technology, Lecture Notes in Electrical Engineering 309, DOI: 10.1007/978-3-642-55038-6_111, © Springer-Verlag Berlin Heidelberg 2014
721
722
J.-R. Chang, H.-C. Juang, and C.-H. Lo
power and apparent power of electrical energy consumption as inputs of the neural network for training and identification [Lam, Fung, and Lee, 2007]. The neural network parameters are optimized by PSO (Particle Swarm Optimization) and Genetic Algorithm to make the best of appliance identification rate for a NILM system. But it needs a smart meter with high frequency sampling to take the transient state. In this study, a smart meter with low frequency sampling (about 1Hz) is used in NILM system. A proposed method takes advantage of the combination of the fuzzy theory and neural network theory automatically identifies the electrical appliance in starting status.
2
Non-Intrusive Load Monitoring System
Nonintrusive Load Monitoring was invented by Hart, Kern and Schweppe of MIT in the early 1980s [Hart, 1989]. NILM system is a process for analyzing changes in the voltage and current going into a house. It deduces what appliances are used in the house as well as their individual energy consumption. The power information is changed dramatically when a new device is plugged in. Different electrical device shows unique electrical characteristic of the power change. It is also known as electr
Data Loading...
