•A stock of buildings is analyzed to discover typical load profiles.•The daily load profiles are grouped with a “Follow the Leader” clustering algorithm.•A globally optimal decision tree is employed ...to develop a customer classifier.•The proposed classifier performs better than the baseline model of about 6%.•The classifier makes use of non-intrusive attributes gathered from energy bills.
The recent increasing spread of Advanced Metering Infrastructure (AMI) has enabled the collection of a huge amount of building related-data which can be exploited by both energy suppliers and users to gain insight on energy consumption patterns. In this context, data analytics-based methodologies can play a key role for performing advanced characterization, benchmarking and classification of buildings according to their typical energy use in the time domain. Traditionally, energy customers are classified according to their building end-use category. However, buildings belonging to the same category can exhibit very different energy patterns making ineffective this kind of a-priori categorization. For this reason, load profiling frameworks have been developed in the last decade to identify homogenous groups of buildings with similar daily energy profiles. The present study proposes a non-intrusive customer classification process, which does not use as predictive attributes in-field load monitoring data for the classification of unknown customers, but rather monthly energy bills and additional information on customers’ habits collected by means of a phone survey. The proposed classification process is developed by analysing hourly energy consumption data of 114 electrical customers of an Italian Energy Provider. The representative daily load profiles are grouped using the “Follow the Leader” clustering algorithm and a globally optimal decision tree is employed to build a supervised classification model. The model, compared to a baseline recursive partitioning tree, leads to an increase of accuracy of about 6%. Eventually, the procedure exploits energy bill data also for estimating the magnitude of typical load profiles.
•The effects of temporal resolutions of energy data are assessed.•The performance of load profiling and forecasting are compared.•Weather impact on load forecasting is evaluated in a univariate and ...multivariate approach.•For data exploration, low-resolution data provides a better indication of significant changes.•In learning mechanism, higher resolution data provides more accurate results.
This study focuses on studying the impact of temporal resolution of energy data from a commercial building on data exploration and learning applications. The goal is to improve the consistency of the modelling techniques and define the best-unified resolution of data for different applications such as data exploration, consistency, load profile extraction, and forecasting for commercial buildings. A three-step process is proposed to evaluate the data resolution effects on mining and learning applications. The first step, the data exploration and consistency, transforms the raw data into a human interpretable form using different temporal features to understand the dependency of the load consumption of the building. In the second step, the K-means clustering technique is used to extract the typical load profiles for all data granularities to deduce information related to the operational behaviour of the building throughout the year. Finally, the long short-term memory model is evaluated for building load forecasting in a univariate and multivariate approach in the last step. The results demonstrate that higher-resolution data does not necessarily assure clear relationships between the operational parameters of a commercial building using data exploration techniques. Furthermore, increasing the granularity of the data did not affect the extracted number of clusters or the load profiles overall shape (peak points). In contrast, to load profile extraction, the obtained results are improved in building load forecasting with NRMSE from 0.125 to 0.028 for daily to 1 min resolution data. Overall, considering the balance of accuracy and processing time the 15-minute resolution data with a univariate approach can perform the best.
The recent technological developments monitoring the electricity use of small customers provides with a whole new view to develop electricity distribution systems, customer-specific services and to ...increase energy efficiency. The analysis of customer load profile and load estimation is an important and popular area of electricity distribution technology and management. In this paper, we present an efficient methodology, based on self-organizing maps (SOM) and clustering methods (K-means and hierarchical clustering), capable of handling large amounts of time-series data in the context of electricity load management research. The proposed methodology was applied on a dataset consisting of hourly measured electricity use data, for 3989 small customers located in Northern-Savo, Finland. Information for the hourly electricity use, for a large numbers of small customers, has been made available only recently. Therefore, this paper presents the first results of making use of these data. The individual customers were classified into user groups based on their electricity use profile. On this basis, new, data-based load curves were calculated for each of these user groups. The new user groups as well as the new-estimated load curves were compared with the existing ones, which were calculated by the electricity company, on the basis of a customer classification scheme and their annual demand for electricity. The index of agreement statistics were used to quantify the agreement between the estimated and observed electricity use. The results indicate that there is a clear improvement when using data-based estimations, while the new-estimated load curves can be utilized directly by existing electricity power systems for more accurate load estimates.
In a competitive retail market, large volumes of smart meter data provide opportunities for load serving entities to enhance their knowledge of customers' electricity consumption behaviors via load ...profiling. Instead of focusing on the shape of the load curves, this paper proposes a novel approach for clustering of electricity consumption behavior dynamics, where "dynamics" refer to transitions and relations between consumption behaviors, or rather consumption levels, in adjacent periods. First, for each individual customer, symbolic aggregate approximation is performed to reduce the scale of the data set, and time-based Markov model is applied to model the dynamic of electricity consumption, transforming the large data set of load curves to several state transition matrixes. Second, a clustering technique by fast search and find of density peaks (CFSFDP) is primarily carried out to obtain the typical dynamics of consumption behavior, with the difference between any two consumption patterns measured by the Kullback-Liebler distance, and to classify the customers into several clusters. To tackle the challenges of big data, the CFSFDP technique is integrated into a divide-and-conquer approach toward big data applications. A numerical case verifies the effectiveness of the proposed models and approaches.
•K-means is used to segment heating consumption intensity and load pattern groups.•A constant load profile represents most of the district heating customers.•Calendar context affects load patterns ...and consumer behavior.•Customers with high energy consumption have lower variability.
The wide use of smart meters enables collection of a large amount of fine-granular time series, which can be used to improve the understanding of consumption behavior and used for consumption optimization. This paper presents a clustering-based knowledge discovery in databases method to analyze residential heating consumption data and evaluate information included in national building databases. The proposed method uses the K-means algorithm to segment consumption groups based on consumption intensity and representative patterns and ranks the groups according to daily consumption. This paper also examines the correlation between energy intensity and the characteristics of buildings and occupants, load profiles of households, consumption behavior changes over time, and consumption variability. The results show that the majority of the customers can be represented by fairly constant load profiles. Calendar context has an impact not only on the patterns but also on the consumption intensity and user behaviors. The variability studies show that consumption patterns are serially correlated, the customers with high energy consumption have lower variability, and the consumption is more stable over time. These findings will be valuable for district heating utilities and energy planners to optimize their operations, design demand-side management strategies, and develop targeting energy-efficiency programs or policies.
The utilization of energy consumption data is crucial for efficient operation and planning in smart grids. Nonetheless, certain obstacles need to be addressed, such as high computational costs, data ...security and privacy concerns, and significant expenses associated with installing smart meters across the electrical grid. To address these challenges, generating synthetic data has emerged as a promising approach, providing an opportunity to enhance energy efficiency, demand flexibility, and power grid operation. Therefore, this study proposes a nonlinear model of independent component estimation (NICE) with convolutional layers to produce realistic load profiles. This research aims to evaluate the potential of deep generative models (DGMs) through the characterization and quantification of electricity consumption profiles obtained from an actual smart grid on a university campus. The Kullback–Leibler divergence is used to evaluate the performance of the proposed model. Simulation results show that the proposed model can accurately capture the spatiotemporal correlation of actual samples, leading to synthetic load profiles that closely resemble actual profiles. The performance of the proposed NICE model is compared with a NICE model with dense layers, as well as with Generative Adversarial Networks (GAN) with dense layers, and GAN with convolutional layers (cGAN), all methods previously used in the literature to generate synthetic load profiles. It was observed that the proposed NICE model with convolutional layers leads to better results. This model produces more significant similarity between the probability distributions of actual and synthetic data, in addition to a more extraordinary ability to reproduce more realistic load variability curves.
•A methodology for generating synthetic load profiles for the smart grids.•Convolutional NICE model for producing realistic load profiles.•Using deep generative models to obtain electricity consumption profiles.•Learning patterns from actual load profiles to reproduce synthetic load profiles.•Generating load profiles for buildings to assist in management and control.
In order to improve low voltage (LV) network visibility without extensive monitoring and integrate low carbon technologies (LCTs) in a cost-effective manner, this paper proposes a novel three-stage ...network load profiling method. It uses real-time information monitored from selective representative areas to develop network templates. The three stages are: clustering, classification and scaling. It can be used to identify the loading conditions of unmonitored LV systems with similar fixed data to those monitored LV substations. In the clustering stage, hierarchical clustering and K-means are used to cluster substations into groups based on the shape of the monitored load profiles. The classification tool designed with multinomial logistic regression maps an unmonitored LV substation into the most probable templates by using routinely available fixed data. Finally, clusterwise weighted constrained regression is employed to estimate peak for individual LV substations and the developed templates. The three-stage profiling is demonstrated on a practical system in the U.K. under the umbrella of a smart grid trail project. Ten LV templates are developed by using the metered data from 800 monitored LV substations. A comprehensive comparison between the estimated peaks using the three-stage process and the metered peaks suggests that the methodology can achieve superior accuracy. This is part I of the paper, introducing clustering and classification. The scaling (peak estimation) process will be introduced in part II of the paper.
•Typical load profile characterization of real Medium Voltage (MV) consumers.•Classification methodology of new MV consumers.•Evaluation of the proposed methodology using several clustering ...algorithms and validity indices.•Data mining techniques to identify load profiles to support the definition of electricity tariffs.
This paper presents an electricity medium voltage (MV) customer characterization framework supported by knowledge discovery in database (KDD). The main idea is to identify typical load profiles (TLP) of MV consumers and to develop a rule set for the automatic classification of new consumers. To achieve our goal a methodology is proposed consisting of several steps: data pre-processing; application of several clustering algorithms to segment the daily load profiles; selection of the best partition, corresponding to the best consumers’ segmentation, based on the assessments of several clustering validity indices; and finally, a classification model is built based on the resulting clusters. To validate the proposed framework, a case study which includes a real database of MV consumers is performed.
A case study of residential electricity consumption patterns mining and abnormal user identification using hierarchical clustering is presented in this paper. First, based on a brief introduction of ...hierarchical clustering, a process model and the specific steps of electricity consumption patterns mining in smart grid environment are proposed. Then, a case study using the daily electricity consumption data of 300 residential users in an eastern city of China, Kunshan, from November 16, 2014 to December 16, 2014, is presented. Through the implementation of hierarchical clustering, 9 abnormal users and 4 types of monthly electricity consumption patterns are successfully identified. The results show that most residential users in Kunshan city, nearly 81%, have a similar monthly electricity consumption pattern. Their average daily electricity consumption is about 7.73 kWh in the early winter with small fluctuations. Also, their daily electricity consumption is significantly associated with the temperature changes. However, it is worth noting that the special electricity consumption patterns of a small proportion of electricity users cannot be ignored, which is of great significance for the planning, operation, policy formulation and decision-making of smart grid.
Demand side management (DSM) is an important strategy for promoting sustainable consumption in resource-rich countries with high purchasing power and subsidized tariffs. Global energy and ...environmental resource consumption have increased rapidly due to advances in production and transportation, leading to inefficient and wasteful use of resources. DSM aims to address these issues by promoting efficiency. To implement appropriate DSM strategies, a greater understanding of consumer behavior is needed. To do so, load profiling and load clustering are two popular methods that can be used. This paper aims to i) summarize the most recent global load profiling and clustering works ii) use official smart meter data to understand key electricity consumption trends in Qatar, such as temperature-demand correlation, weekend vs. weekday, and public holiday consumption patterns, and iii) perform load clustering to propose policies that would help manage the electricity load in Qatar for its green growth. This study provides insights into the electricity consumption trends of various sectors in Qatar, including commercial, government, hospitality, and residential sectors. It was found that among all the sectors there were only two usage periods of the same time. There is naturally a strong correlation between temperature and electricity consumption throughout the sectors. Furthermore, it was observed that the consumption of the sectors is highly similar which leads to multiple sectors being present in the same cluster. Finally, policy changes are proposed based on the results to encourage demand response programs in Qatar.