Ali Asheibi1, Zakariya Rajab2 and Ashraf Khalil3, 1Department of Electronics and Communications Engineering, A’Sharqiyah University, Oman, 2Department of Electrical and Electronic Engineering, University of Benghazi, Libya, 3DTU Engineering Technology, Technical University of Denmark, Denmark
The increasing penetration of clean energies play challenges to power quality due to their power output variability because of their dependency on weather conditions. Therefore, the need for continuous power quality monitoring becomes very essential to capture any power quality (PQ) disturbance events in relation to unusual weather conditions. In this paper, clustering as an unsupervised learning technique is used to study the relationship between PQ disturbances events and weather data. The data is modelled by Gaussian Mixture Models (GMM) which is a technique uses a mixture of Gaussian distributions using the mean vector and covariance matrix. To train the GMM, the algorithm of expectation and maximization (EM) is employed. Gaussian mixture models (GMM) Clustering method is applied for PQ events and weather data sets, to discover any association between the PQ events and the weather conditions. The Supervised learning using C5.0 algorithm is then applied to the discovered classes to gain close insight into the obtained clusters and to predict the occurrences of unusual clusters in future measurement data.
Power Quality (PQ), Clustering, Classification, Gaussian Mixture Models (GMM), Decision Trees (DT).
Wei Zhang, Ziyu Chen, Jiapeng Cai and Yingying Li, Department of Electrical and Computer Engineering, Nanfang College Guangzhou, Guangzhou, China
With the rapid development of high-rise buildings, the application of elevator is increasing day by day. In order to solve the problem of more reasonable and efficient operation of elevator in the peak of passenger flow, this paper identifies the characteristics of passenger flow by counting the number of up calls and down calls in a period of time, which provides a basis for the optimization of elevator scheduling and coordination control. In this thesis, the perceptron model in machine learning is used to realize the elevator traffic pattern recognition. Through the training of the existing data to construct the traffic pattern recognition model, and then through the recognition of the model to verify the test data, finally we can achieve the correct identification of the elevator traffic pattern.
Machine learning, Elevator, Pattern recognition.
Copyright © ELEG 2022