Power Usage Patterns Research Articles

Utilizing Machine Learning Techniques for Power Estimation in Residential Electricity Consumption

Electricity consumption analysis and estimation play a vital role in various contexts, ranging from business planning to household financial management. In an era where electrical energy is a primary source of human activity, understanding power usage patterns and estimating them becomes crucial. This research enables an understanding of energy consumption from individual to national scales and supports efficient energy policy planning. Power estimation provides insights into the expenditures that consumers or companies will incur, enabling effective financial planning. Research and development of accurate predictive models using machine learning techniques are essential to providing useful information for the energy industry and the general public. This study uses machine learning to compare the performance of different power prediction models. The decision tree (DT) method, with an MAE of 0.3613, an MSE of 0.4184, and an RMSE of 0.6469, gives the best results. The results show that the model can provide accurate predictions of electricity consumption, potentially becoming a reliable tool in power management and financial management. Its contribution is crucial in the energy industry and daily life, providing insights needed to enhance energy efficiency and budget management. As a result, this research provides relevant and beneficial solutions to energy and financial management for society and industry. Keywords: Electricity consumption analysis, Estimation, Machine learning, Power management, financial management

Perfednilm: a practical personalized federated learning-based non-intrusive load monitoring

Non-Intrusive Load Monitoring (NILM) is a valuable technique for breaking down overall power consumption into the energy usage of individual appliances. Understanding power usage patterns through NILM plays an important role in reducing energy costs and achieving carbon reduction goals. However, privacy concerns often deter consumers from sharing their electricity consumption data. To address these privacy concerns, Federated Learning (FL) has been introduced in NILM, which enables the training of NILM models while keeping power consumers’ data locally. However, FL’s reliance on a single global model leads to poor performance on clients with unique power consumption patterns. In response to this challenge, we present a Personalized Federated Learning NILM algorithm (PerFedNILM), a practical personalized FL approach for NILM. PerFedNILM limits the local update bias across clients and trains personalized models for individual clients to improve load-monitoring performance. In addition, it mitigates the negative impact of client dropout, which is a common issue in practice. Our experiments on using real-world energy data demonstrate that PerFedNILM outperforms previous FL-based NILM methods, especially in client dropout scenarios.

Vertical Transportation System Power Usage: Behavioural Case Study of Regulated Buildings in Bangkok

Sustainable urban development worldwide is crucial for the development of living spaces in high-rise buildings and infrastructures, which leads to the inevitability of increased energy consumption and demand of vertical transportation systems. The evaluation of the energy consumption of transportation systems is needed to verify and analyse the power usage related to traffic demands and patterns. In addition, efficient vertical transportation systems are central to the formulation of more sustainable cities. Therefore, this trend represents a substantial portion of the overall energy consumption of the building types. The benchmarking of the energy needs of the vertical transportation systems in five different building types via the comparison of granular load profile patterns (in conjunction with population densities) to the energy consumed was conducted, and it will be used to infer some impactful design strategies for the future. This study demonstrated a systematic approach to determine the power usage patterns in vertical transportation systems by actual measurement and traffic data collection from elevator monitoring. This may be used to develop a prediction for other cases in different types of installed vertical transportation systems. Therefore, the power usage of the vertical transportation systems can be used to determine the correlation between energy consumption and load pattern based on building characteristics and the overall energy consumption of each presented system.

Energy Management Solution using Internet of things

Energy management is imperative at a time like this. Due to rising population and lack of adequate power supply to satisfy energy consumers, there is an urgent need to provide an efficient way to manage energy. This study proffers solution to the enormous challenge of managing energy by developing a model that helps both consumer and energy providers manage energy efficiently. This solution will enable monitoring of power over a specific period and intelligent switching of power sources. Thus, the energy consumer gets a detailed information of energy usage and specific time and duration of power switching between different energy sources. It assists energy providers such as the Electricity DISCOs (Distribution Companies) to monitor energy distribution. The study composes four distinct parts which are; the automatic transfer switch, the internet of things module, the mobile application and the cloud server. This study helps to intelligently chose between energy sources, observe and log power usage patterns of homes, provide analytics and insight to energy usage by consumers, provide ease, comfort and safety to users and optimally manage energy in homes.

Nonintrusive Load Monitoring Using an LSTM With Feedback Structure

Many non-intrusive load monitoring (NILM) studies use high-frequency data to classify the device’s ON/OFF state. However, these approaches cannot be applied in real-world situations due to increased network traffic and database capacity issues. For these reasons, when trying to perform NILM with low-frequency data, the power usage pattern that changes over time disappears and features cannot be properly obtained to classify devices. In this article, we propose a novel NILM model that can learn datasets with imbalanced data classes. The model extracts features using long-short term memory (LSTM) and improve the feature representation ability of LSTM through the feedback of predictions. The experiment is conducted using the REDD dataset and the Living-lab validation dataset. In the REDD dataset, the proposed method outperforms conventional methods 10%–20% on the Majority device and 50%–60% on the Minority device. Living-lab validation results show that the performance of the proposed method outperforms other previously proposed NILM systems in low-frequency data and can be applied to real-world NILM situations.

Application TadGAN to Detect Collective Anomaly in Power Usage Data

In order to develop an efficient power generation plan, it is necessary to identify consumers’ power usage patterns. In general, power usage data takes the form of time series data and in order to analyze that data, it is necessary to first verify that there is no data contamination. To this end, the process of verifying that there are no anomalies in the data is essential. In particular, for power data, anomalies are often recorded across multiple time units rather than just one point. In this work, we applied the TadGAN algorithm to detect these collective anomalies. Using the power usage data recorded in the actual building, the anomalies were injected randomly with various conditions. Afterwards, we detected anomalies by using TadGAN and showed that we were better at detecting collective anomaly than point anomaly.

Household power usage pattern filtering-based residential electricity plan recommender system

Deregulation of the retail electricity market has led to the emergence of an increasing number of electricity plans with competitive rates. Electricity customers now have more flexibility in choosing an electricity provider and electricity plan based on individual consumption needs. In this paper, a feature engineering hybrid collaborative filtering-based electricity plan recommender system (FECF-EPRS) is proposed for helping the customer get the right electricity plan. This system is composed of three-segment models for missing feature estimation, feature crosses construction, and electricity plan recommendation. It only takes easy-to-obtain household appliance usage features as inputs and outputs ratings for different plans. Through the test of real electricity market data, the FECF-EPRS shows a greater improvement in terms of recommendation accuracy, which can provide more accurate recommendations to customers and more reasonable pricing references for retailers.

An Optimization Based Power Usage Scheduling Strategy Using Photovoltaic-Battery System for Demand-Side Management in Smart Grid

Due to rapid population growth, technology, and economic development, electricity demand is rising, causing a gap between energy production and demand. With the emergence of the smart grid, residents can schedule their energy usage in response to the Demand Response (DR) program offered by a utility company to cope with the gap between demand and supply. This work first proposes a novel optimization-based energy management framework that adapts consumer power usage patterns using real-time pricing signals and generation from utility and photovoltaic-battery systems to minimize electricity cost, to reduce carbon emission, and to mitigate peak power consumption subjected to alleviating rebound peak generation. Secondly, a Hybrid Genetic Ant Colony Optimization (HGACO) algorithm is proposed to solve the complete scheduling model for three scenarios: without photovoltaic-battery systems, with photovoltaic systems, and with photovoltaic-battery systems. Thirdly, rebound peak generation is restricted by using Multiple Knapsack Problem (MKP) in the proposed algorithm. The presented model reduces the cost of using electricity, alleviates the peak load and peak-valley, mitigates carbon emission, and avoids rebound peaks without posing high discomfort to the consumers. To evaluate the applicability of the proposed framework comparatively with existing frameworks, simulations are conducted. The results show that the proposed HGACO algorithm reduced electricity cost, carbon emission, and peak load by 49.51%, 48.01%, and 25.72% in scenario I; by 55.85%, 54.22%, and 21.69% in scenario II, and by 59.06%, 57.42%, and 17.40% in scenario III, respectively, compared to without scheduling. Thus, the proposed HGACO algorithm-based energy management framework outperforms existing frameworks based on Ant Colony Optimization (ACO) algorithm, Particle Swarm Optimization (PSO) algorithm, Genetic Algorithm (GA), Hybrid Genetic Particle swarm Optimization (HGPO) algorithm.

Analytical Program Power Characterization for Battery Depletion-time Estimation

Appropriate battery selection is a major design decision regarding the fast growth of battery-operated devices like space rovers, wireless sensor network nodes, rescue robots, and so on. Many such systems are mission critical, where estimation of the battery depletion time has an important role in the design efficiency with regard to the mission time. Accurate characterization of the system power usage pattern is essential for such an estimation. The following complexities exist: (1) The system behavior changes during interaction with the physical world, (2) the power consumption varies as the runtime progresses, (3) the total delivered battery charge has non-linear dependency on the power variability, and (4) design-time exhaustive study about runtime execution paths is almost impossible. This article presents an analytical method to first characterize the power variability of a given embedded program modeled by a directed acyclic graph, concerning the first and the second complexities. To include the third complexity, however, the concept of Worst-case Power Consumption Trace (WPCT) is proposed toward the worst-case scenario in terms of charge depletion for a given battery. A polynomial algorithm is also presented to construct WPCT and use it to estimate a tight lower bound for the system energy depletion time, i.e., its failure time, avoiding an exhaustive study. Comparisons between the analytical and simulation results reveal less than 3.4% of error in the bound estimations for the considered setups.

Approximation Algorithm-Based Prosumer Scheduling for Microgrids

Since the inherent intermittency and uncertainty of renewable energy resources complicates efficient Microgrid operations, a Demand Response (DR) scheme is implemented for customers in the grid to alter their power-usage patterns. However, for a real-time pricing model at the current DR, the automated decision on the energy price is not trustworthy because of artificial interferences to the power generation. As opposed to energy price, an operational cost-based prosumer scheduling approach would be able to protect the integrity of the power grid operations from deceptive market transactions and assist in robust energy management. To investigate the operational challenges associated with the costs and prosumers in the Microgrid, we focus on formulating the problem mathematically and designing approximation algorithms to solve the problem of how to optimally identify suppliers to minimize the total operational costs associated with providing electricity. We prove the hardness of the scheduling as one of the NP-Hard problems and propose polynomial time algorithms for approximating optimal solutions. With a proper resilience level for reliable power services, the scheduling algorithms include ways to construct not only robust supplier networks, but also group energy communities in terms of black start while minimizing the operational costs. A series of theoretical performances and experimental evaluations also demonstrates the practical effectiveness of this scheduling model for the operations.

Demand-Side Management of Solar Microgrid Operation: Effect of Time-of-Use Pricing and Incentives

Over 17% of the world’s population lack access to electricity, the majority being in rural areas of sub-Saharan Africa and South Asia. Microgrid technologies are a promising solution towards rural and remote area electrification; however, ever-increasing electricity demand remains a big challenge leading to pronounced power outages. Demand-side management is an indispensable tool towards addressing the challenges. This paper employs a mathematical model based on incentives and time-of-use rates to simulate daily power usage pattern of residential customers using data collected from an isolated village Ngurdoto solar microgrid, Arusha, Tanzania. Customer responsiveness on the increase in price was evaluated based on the concept of price elasticity of demand. Using two demand response strategies, namely, load shifting (LS) and scheduled load reduction (SLR), the results reveal that LS can achieve up to 4.87% energy-saving, 19.23% cost-saving, and about 31% and 19% peak reduction and power factor improvement, respectively. SLR method resulted in about 19% energy-saving, 49% cost-saving, and 24% power factor improvement. Thus, the results presented in this study may lead to a more efficient and stable system than the current state in developing countries’ utility.

A Cluster-Based Baseline Load Calculation Approach for Individual Industrial and Commercial Customer

Demand response (DR) in the wholesale electricity market provides an economical and efficient way for customers to participate in the trade during the DR event period. There are various methods to measure the performance of a DR program, among which customer baseline load (CBL) is the most important method in this regard. It provides a prediction of counterfactual consumption levels that customer load would have been without a DR program. Actually, it is an expected load profile. Since the calculation of CBL should be fair and simple, the typical methods that are based on the average model and regression model are the two widely used methods. In this paper, a cluster-based approach is proposed considering the multiple power usage patterns of an individual customer throughout the year. It divides loads of a customer into different types of power usage patterns and it implicitly incorporates the impact of weather and holiday into the CBL calculation. As a result, different baseline calculation approaches could be applied to each customer according to the type of his power usage patterns. Finally, several case studies are conducted on the actual utility meter data, through which the effectiveness of the proposed CBL calculation approach is verified.

Multiclass classification of mobile applications as per energy consumption

Today even the commonly used smartphones have high computational power and storage capacity, which make even execution of resource intensive operations possible with them. The increase in computational capability and varied communication features, have resulted in enhancement of energy requirement of these devices. The increasing popularity of mobile computing is always bottlenecked with its battery life. To study the power consumption pattern experiment was conducted on some popular mobile applications under different environment. The energy used by applications was recorded and used as training set to develop a machine learning model that classify these mobiles applications into three categories namely low, medium and high depending on its energy usage, using classification approach. Various Multiclass classification algorithms are explored to find the most accurate model that can be used for classifying the mobile applications. Every day a new mobile application is created for catering a wide range of applications. These applications have new and more interactive features, and consume relatively high energy for execution, leading to drainage of battery. The detailed analysis of logical reasoning behind machine learning models developed in this study will provide designer deeper insight of type of mobile application and its power usage pattern.

Demand response of customers in Kitakyushu smart community project to critical peak pricing of electricity

Dynamic pricing is an effective way to induce the customers to change their electricity usage patterns, critical peak pricing (CPP) event shows a great potential to reduce the peak demand and enhance the grid reliability. This paper investigates performances of the CPP experiment across aggregated group-level of electricity customers covering 16 commercial, 15 office and 50 residential objectives in the Kitakyushu Smart Community demonstration project. A data-based approach to the comparisons of electricity demand curves and distributions among time of use (TOU) and CPP events. Result confirms the differences in price-based responsiveness across types of customer, there is demand shaping effect during the CPP event period by adjusting the electricity price at 30-min interval. The power saving effects were obviously confirmed in residential and commercial sectors under CPP event program. A multiple linear regression model is used to examine the response effectiveness and weigh the influencing factors of the fixed time for the dynamic pricing scheme. For the no-residential part, the electricity consumption strongly depends on the variations in temperature, commercial consumer tends to respond more to the changes of electricity price than office sector. The residential customer shows a promising potential for load reduction corresponding the increasing electricity price. However, the effects depend on the preferences of residential customers to change their power usage pattern at real time of the CPP event period.

COMMUNITY BASED HOME ENERGY MANAGEMENT SYSTEM

In a Smart Grid (SG) scenario, domestic consumers can gain cost reduction benefit by scheduling their Appliance Activation Time (AAT) towards the slots of low charge. Minimization in cost is essential in Home Energy Management Systems (HEMS) to induce consumers acceptance for power scheduling to accommodate for a Demand Response (DR) at peak hours. Despite the fact that many algorithms address the power scheduling for HEMS, community based optimization has not been the focus. This paper presents an algorithm that targets the minimization of energy costs of whole community while keeping a low Peak to Average Ratio (PAR) and smooth Power Usage Pattern (PUP). Objective of cost reduction is accomplished by finding most favorable AAT by Particle Swarm Optimization (PSO) in conjunction with Inclined Block Rate (IBR) approach and Circular Price Shift (CPS). Simulated numerical results demonstrate the effectiveness of CPS to assist the merger of PSO & IBR to enhance the reduction/stability of PAR and cost reduction.

Two-Stage Load Pattern Clustering Using Fast Wavelet Transformation

Smart grids collect large volumes of smart meter data in the form of time series, or so-called load patterns. We outline the applications that benefit from analyzing this data (ranging from customer segmentation to operational system planning), and propose two-stage load pattern clustering. The first stage is performed per individual user and identifies the various typical daily power usage patterns (s)he exhibits. The second stage takes those typical user patterns as input to group users that are similar. To improve scalability, we use fast wavelet transformation (FWT) of the time series data, which reduces the dimensionality of the feature space where the clustering algorithm operates (i.e., from ${N}$ data points in the time domain to log ${N}$ ). Another qualitative benefit of FWT is that patterns that are identical in shape, but just differ in a (typically small) time shift still end up in the same cluster. Furthermore, we use ${g}$ -means instead of ${k}$ -means as the clustering algorithm. Our comprehensive set of experiments analyzes the impact of using FWT versus time-domain features, and ${g}$ - versus ${k}$ -means, to conclude that in terms of cluster quality metrics our system is comparable to state-of-the-art methods, while being more scalable (because of the dimensionality reduction).

Fine-grained appliance usage and energy monitoring through mobile and power-line sensing

To promote energy-efficient operations in residential and office buildings, non-intrusive load monitoring (NILM) techniques have been proposed to infer the fine-grained power consumption and usage patterns of appliances from power-line measurement data. Fine-grained monitoring of everyday appliances (such as toasters and coffee makers) can not only promote energy-efficient building operations, but also provide unique insights into the context and activities of individuals. Current building-level NILM techniques are unable to identify the consumption characteristics of relatively low-load appliances, whereas smart-plug based solutions incur significant deployment and maintenance costs. In this paper, we investigate an intermediate architecture, where smart circuit breakers provide measurements of aggregate power consumption at room (or section) level granularity. We then investigate techniques to identify the usage and energy consumption of individual appliances from such measurements. We first develop a novel correlation-based approach called CBPA to identify individual appliances based on both their unique transient and steady-state power signatures. While promising, CBPA fails when the set of candidate appliances is too large. To further improve the accuracy of appliance level usage estimation, we then propose a hybrid system called AARPA, which uses mobile sensing to first infer high-level activities of daily living (ADLs), and then uses knowledge of such ADLs to effectively reduce the set of candidate appliances that potentially contribute to the aggregate readings at any point. We evaluate two variants of this algorithm, and show, using real-life data traces gathered from 10 domestic users, that our fusion of mobile and power-line sensing is very promising: it identified all devices that were used in each data trace, and it identified the usage duration and energy consumption of low-load consumer appliances with ∼87% accuracy.

The Privacy Analysis of Battery Control Mechanisms in Demand Response: Revealing State Approach and Rate Distortion Bounds

Demand response systems in the electricity grid, which rely on two way communication between the consumers and utility, require the transmission of instantaneous energy consumption to utilities. Perfect knowledge of a user’s power consumption profile by a utility is a violation of privacy and can be detrimental to the successful implementation of demand response systems. It has been shown that an in-home energy storage system (such as a battery/inverter) that provides a viable means to achieve the cost savings of instantaneous electricity pricing without inconvenience can also be used to hide a user’s power usage pattern. A fundamental tradeoff exists between the costs saved and the degree of privacy achievable, and in this paper, the tradeoff achievable by a finite capacity battery assuming a zero tolerance for activity delay is studied using a Markov process model for user’s demands and instantaneous electricity prices. Due to high computational complexity (continuous state-action space) of the stochastic control model, inner and upper bounds are presented on the optimal tradeoff. In particular, a class of battery charging policies based on minimizing revealing states is proposed to derive achievable privacy-cost savings tradeoff. The performance of this algorithm is compared with lower bounds derived using a greedy heuristic and upper bounds derived using an information theoretic rate distortion approach. The framework proposed is shown to be applicable even when users only desire partial information protection, such as presence/absence of activity or specific appliances they wish to hide. Numerical results based on real electricity and pricing data show that the proposed algorithm performs close to the upper bound demonstrating its efficacy.

Coordinated energy management for inter-community imbalance minimization

With the increase in demand flexibility and the rapid introduction of uncontrollable renewable power sources, effective schemes for managing the power usage of end-users are required. In this line, we propose an energy management framework for coordinating the power usage of communities of networked agents. More specifically, communities (groups of loads) coordinate to minimize their aggregated power imbalance, while taking into account each community's objectives and constraints, as well as the preferred power usage pattern of each end-user. For having a robust coordination that can work under unexpected events, we propose to assign the agents to communities using a measure of the flexibility of sets of agents. The coordination framework builds on the alternating directions method of multipliers (ADMM), algorithm that is used to implement a distributed coordination using a hierarchical architecture. While the distributed coordination allows to manage the power usage of each end-user, the hierarchical architecture enables the integration, in a single framework, of energy management problems that would be otherwise handled independently. We illustrate and analyze the coordination framework using three simulated scenarios.

Power Consumption Based Cloud Scheduler

After years of development and trial usage, cloud computing has finally entered into mainstream and wide spread commercial use. Among the various cost elements in operating a cloud computing infrastructure, power usage stands out prominently. In fact, how to reduce the power consumption of cloud computing platforms has received considerable attention of the cloud service provider as well as academic researchers. This paper proposes a power awareness scheduler (PAS) system to manage the resource scheduling based on power usage patterns in the cloud. By obtaining and analyzing the power consumption-load curve, PAS is able to make power-efficient virtual machine deployment for the cloud. Our implementation of the proposed PAS on OpenStack shows that PAS can reduce total power consumption by up to 17%.