Non-intrusive load monitoring (NILM) is an effective method to optimize energy consumption patterns. Since the concept of NILM was proposed, extensive research has focused on energy disaggregation or load identification. The traditional method is to disaggregate mixed signals, and then identify the independent load. This paper proposes a multi-label classification method using Random Forest (RF) as a learning algorithm for non-intrusive load identification. Multi-label classification can be used to determine which categories data belong to. This classification can help to identify the operation states of independent loads from mixed signals without disaggregation. The experiments are conducted in real environment and public data set respectively. Several basic electrical features are selected as the classification feature to build the classification model. These features are also compared to select the most suitable features for classification by feature importance parameters. The classification accuracy and F-score of the proposed method can reach 0.97 and 0.98, respectively.

The Internet of Things (IoT) connects electric devices to a network in order to realize virtual control. Demand-side loads can provide ancillary services that optimize energy structures and promote supply–demand balance. For load management, this article proposes Internet of Electric Loads (IoEL) as a service-centric network, and the edge subcloud is introduced in the distributed computing system. The intelligent thermostatically controlled loads constitute the physical perceptual layer of the IoEL. Each subcloud controls a type of loads under the same regional distribution grid to form a stable and controllable aggregation system. The subclouds interact with the load cloud to form a platform layer. Through the interaction between the platform and physical perceptual layers, loads are controlled to realize the load transfer and provide ancillary services for the application layer. IoEL utilizes computing resources from different layers to form a distributed computing system made up of cloud, edge, and end devices. The load management is divided into four steps, namely, load aggregation in the edge–end interface, evaluation of load capacity in the edge, task allocation in the cloud–edge interface, and load control in the edge–end interface. The loads upload the physical properties to the subclouds and are then aggregated according to similarities of properties. The subclouds then evaluate schedulable capacity and upload these evaluations to the cloud, which then allocates tasks to each subcloud accordingly. The edge is responsible for receiving tasks from the cloud and controlling loads. The effectiveness of the method was verified by providing clean energy output tracking.

Electric vehicles (EVs) with both power consumption and energy storage are important load resources involved in the adjustment of supply and demand. The aggregation and control of a large number of EVs is a way to make the energy structure upgrade. This paper plans to explore the cluster response mechanism, group effect and coordination ability of distributed EVs in the regional power grid on the basis of solving the following problems. Firstly, the EV cluster presents heterogeneous characteristics due to differences of physical parameters. Secondly, EVs participating in the aggregation dynamically change in real time because of the stochastic charging behaviors. In this paper, a heterogeneous aggregation model for EVs with random charging behaviors is proposed based on the changing relation of charging power to remaining electric quantity. And a variable sliding mode control model is constructed to address the randomness of the charging process and realize stable responses in a short-time scale. Besides, the battery module and the user strategy are proposed to coordinate with the heterogeneous cluster to improve the adjustment capacity and flexibility. Finally, A clean energy output tracking simulation is provided to verify the effectiveness of the heterogeneous aggregation and control model.

Non-intrusive load monitoring (NILM) is performed to realize intelligent power consumption. A load identification algorithm which is flexible for various households is required to realize automatic NILM. An approach of NILM proposed in this paper is general for most types of loads and exclusive for different users. In this paper, inspired by the knowledge graph holding strong power for knowledge representing, a load feature graph that preserves sufficient information is constructed. The graph provides a net relation between load types and features, which is the basis of load knowledge reasoning. After transforming current and voltage into structured images and data according to the net relation, the convolutional neural networks and combined Support Vector Machines are used to classify images and data respectively. Furthermore, to minimize running time, a customized load identity library focusing on the individual user is proposed. Considering requirements of different computing power resources, above processes are deployed on the architecture of end-cloud collaboration. The experimental results on NILM datasets demonstrate that the proposed method outperforms most of the existing methods in the recognition accuracy and supports for different types of users.

Non-intrusive load monitoring (NILM) is a cost-effective technique for extracting device-level energy consumption information by monitoring the aggregated signal at the entrance of the electric power. With the large-scale deployment of smart metering, NILM should ideally be designed to operate purely on the low-rate data from smart meters. In this paper, an approach based on Graph Shift Quadratic Form constrained Active Power Disaggregation (GSQF-APD) is proposed, which is built upon matrix factorization and introduces graph shift quadratic form constraint according to piecewise smoothness of the power signal. In addition, a two-step iterative optimization method is designed to solve this problem. The first step minimizes the regularization term to find the signal with minimum variation, and then the second step uses the simulated annealing (SA) algorithm to iteratively minimize the objective function and constraint based on the total graph variation minimizer. Using one open-access dataset, the strength of GSQF-APD is demonstrated through three sets of experiments. The numerical results show the superior performance of GSQF-APD, with Graph Laplacian Quadratic Form constrained Active Power Disaggregation (GLQF-APD) and the state-of-the-art NILM methods as benchmarks.

Non-intrusive load monitoring (NILM) is a core technology for demand response (DR) and energy conservation services. Traditional NILM methods are rarely combined with practical applications, and most studies aim to disaggregate the whole loads in a household, which leads to low identification accuracy. In this method, the event detection method is used to obtain the switching event sets of all loads, and the power consumption curves of independent unknown electrical appliances in a period are disaggregated by utilizing comprehensive features. A linear discriminant classifier group based on multi-feature global similarity is used for load identification. The uniqueness of our algorithm is that it designs an event detector based on steady-state segmentation and a linear discriminant classifier group based on multi-feature global similarity. The simulation is carried out on an open source data set. The results demonstrate the effectiveness and high accuracy of the multi-feature integrated classification (MFIC) algorithm by using the state-of-the-art NILM methods as benchmarks.