סמינר מקוון עם אמיתי קליגמן

קישור לזום

אבסטרקט

Non-Intrusive Load Monitoring (NILM) process is intended for separating individual power features from an aggregated energy reading, in order to estimate the operation of individual appliances.

In the past, electricity meters specified only active power reading, for billing purposes, limiting the abilities of NILM solutions. However, recent progress in smart metering technology introduced cost-effective, household consumer-grade metering products, which can produce multiple features with high accuracy.

In this research, a new technique is proposed to apply BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies) method as part of a multi-dimensional load disaggregation solution, based on extraction of multiple features from a smart meter.

The new technique is called Disaggregation in NILM by means of multidimensional BIRCH clustering (DNB).

The main contribution of this research is a technique, comprises four-steps for NILM solution that utilizes a multi-dimensional feature-space with adaptation to power quality parameters.

The proposed algorithm is simple, fast, light-weight, can use raw data samples from a smart meter, can use low-frequency samples reading and does not require a preliminary training or learning process. The proposed method was tested using a private dataset, demonstrating a good use of multi-dimensional space. Further tests involved a limited-dimensional scenario, using AMPds public dataset, in order to compare performance to other NILM algorithms. Some additional experiments in a severe-noise environment were also performed, in order to further explore DNB boundaries and limitations.