Search

Electrical disaggregation, also known as non-intrusive load monitoring (NILM) or non-intrusive appliance load monitoring (NIALM), attempts to recognize the energy consumption of single electrical appliances from the aggregated signal. This capability unlocks several applications, such as giving feedback to users regarding their energy consumption patterns or helping distribution system operators (DSOs) to recognize loads which could be shifted to stabilize the electrical grid. The project “SmartNIALMeter” brought together universities, companies and DSOs and involved the collection of a large data corpus comprising 20 buildings with a total of 100 electrical appliances for a period of up to two years at a sampling interval of five seconds. The variability of the loads, including heat pumps and a charging station for electric vehicles, and the presence of single-phase and three-phase devices make this dataset suitable for several investigations. The total consumption was collected through smart meters and each appliance's consumption was measured with a dedicated sensor, providing sub-metering for all loads. The dataset can be used to tackle several open research questions, for example to investigate new NILM algorithms able to learn with a limited amount of sub-metered data.

Optimal behind-the-meter energy management often requires a day-ahead electric load forecast capable of learning non-linear and non-stationary patterns, due to the spatial disaggregation of loads and concept drift associated with time-varying physics and behavior. There are many promising machine learning techniques in the literature, but black box models lack explainability and therefore confidence in the models’ robustness can’t be achieved without thorough testing on data sets with varying and representative statistical properties. Therefore this work adopts and builds on some of the highest-performing load forecasting tools in the literature, which are Long Short-Term Memory recurrent networks, Empirical Mode Decomposition for feature engineering, and k-means clustering for outlier detection, and tests a combined methodology on seven different load data sets from six different load sectors. Forecast test set results are benchmarked against a seasonal naive model and SARIMA. The resultant skill scores range from −6.3% to 73%, indicating that the methodology adopted is often but not exclusively effective relative to the benchmarks.

The Nigerian Universities rely on weak and unreliable fossil-based electric grids with diesel engine generators (DEG) as a backup. However, there is a potential to light up the campuses using power systems derived from primary renewable power systems (RPS) like wind turbine (WT) and solar photovoltaic (PV), that can be on or off-grid to improve the energy mix and duration reliably. This study presents the comparative analysis of the optimal hybrid grid and off-grid systems (OGS & OOGS) for serving the demand load of university buildings in four climatic regions of Nigeria. HOMER Pro is used to design and select the systems based on minimal net present cost (NPC) and cost of electricity (COE). The impact of a minimal renewable fraction of 95% on the optimal system architecture (OSA) and COE is studied for both grid and off-grid modes. Also, sensitivity analysis of the impact of key variables on performance for the sites is carried out. It is found that the OGS in the four regions is PV/Converter (Conv), while for the OOGS, it is PV/WT/DEG/battery (BB)/Conv except in Port Harcourt (PH), where it is PV/DEG/BB/Conv. The COE for the OGS in the Savana and monsoon climes of Enugu and PH are 10 and 19% more than that in the warm-semi arid climate zones of Maiduguri and Kano, which is approximately 0.09 $/kWh. The COE ($/kWh) for the OOGS is 0.21 in Maiduguri, 0.245 in Kano, 0.275 in Enugu and 0.338 in PH. An obligatory 95% RF changes the architecture and increases COE in all the locations except Maiduguri, with a slightly improved COE but higher NPC like other locations. It is established that the suggested hybrid system is beneficial and feasible for supplying more reliable and clean energy to educational buildings in different Nigerian locations.

Parallel operation by multiple power converters is typical in microgrids. The communication method for parallel control mainly applies MODBUS RTU (Remote Terminal Unit), and the control performance largely depends on the communication speed. This paper proposes a new communication method that enables fast and accurate communication between master and slave systems combined in parallel using the PWM and capture functions of DSP used as the power converter controller. In the proposed communication method, the transmitting side uses the PWM function of DSP to generate a pulse waveform whose pulse width changes according to the reference command and transmits it to the receiving side. Suppose the capture function of the receiving side DSP is set to the time difference rising and falling edge mode. The set event occurs when the capture signal is detected and initializes the counter value to zero. At this time, it detects the counter signal and receives data. We explain the proposed communication method’s principle, characteristics, and noise-canceling method and evaluate the accuracy and stability through data transmission and reception experiments. Finally, three 10kW single-phase electronic loads are connected to the R, S, and T phases to form a three-phase electronic load, and a master and two slaves are designated. The validity of the proposed communication method is verified through the current control response according to the command.

Renewable energy production has been surging around the world in recent years. To mitigate the increasing uncertainty and intermittency of the renewable generation, proactive demand response algorithms and programs are proposed and developed to further improve the utilization of load flexibility and increase the efficiency of power system operation. One of the biggest challenges to efficient control and operation of demand response resources is how to forecast the baseline electricity consumption and estimate the load impact from demand response resources accurately. In this paper, we propose a mixed effect segmented regression model and a new robust estimate for forecasting the baseline electricity consumption in Southern California, USA, by combining the ideas of random effect regression model, segmented regression model, and the least trimmed squares estimate. Since the log-likelihood of the considered model is not differentiable at breakpoints, we propose a new backfitting algorithm to estimate the unknown parameters. The estimation performance of the new estimation procedure has been demonstrated with both simulation studies and the real data application for the electric load baseline forecasting in Southern California.

Aiming at the nonlinear characteristics of short-term electric load series, a short-term electric load forecasting model with improved Extreme Learning Machine (ELM) is proposed. This model is based on the data feature analysis of Ensemble Empirical Mode Decomposition (EEMD) and Long Short-Term Memory (LSTM). EEMD is used to decompose the load sequence into several regular modal components to reduce the error caused by the randomness of the sequence. LSTM and ELM are used to predict the high-frequency and low-frequency components respectively. After using the improved Extreme Learning Machine (ELM), the prediction results of each component are solved. Then the optimal results of power load forecasting are obtained. The results show that the proposed model can judge the overall change trend of the electric load and predict the details of the local change. The average absolute percentage error is 2.24% comparing to 4.78% of the best experiment result of single model, and it has better generality.

Based on the performed researches the method of decomposition of graphs of total electric loading of power system with application of a method of Hilbert-Huang is improved. This approach allows obtaining a homogeneous basic com-ponent of electrical load and temperature component, which has a close correlation with air temperature, which im-proves the accuracy of short-term forecasting. The results of testing the developed mathematical model are given. Ref. 9, figure, table.

Accurate forecasting of an electric load is vital in the effective management of a power system, especially in flourishing regions. A new hybrid model called logarithmic spiral firefly algorithm-support vector regression (LS-FA-SVR) is proposed to promote the performance of electric load forecasting. The new hybrid model is acquired by combining the support vector regression, firefly algorithm, and logarithmic spiral. Half-hourly electric load from five main regions (NSW, QLD, SA, TAS, and VIC) of Australia are used to train and test the proposed model. By comparing the model results with observed data on the basis of the root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percent error (MAPE), the performance of the proposed hybrid model is the most outstanding among all the considered benchmark models. Hence, the results of this study show that the hybrid model LS-FA-SVR is preferable and can be applied successfully because of its high accuracy.

The article concerns design of electric system with dispersed generation, more precisely, functional location of energy generation units — transmission losses relationship. Electrical load centers are the optimal locations of energy generation units, but time-varying loads shift them and transmission losses increase. The purpose of the article is development of an algorithm and the program of power supplies dynamic redistribution’s for the electric system with dispersed generation that reduce transmission losses. Power supplies dynamic redistribution’s is achieved by the creation of a virtual electric system with the possibility of redistribution over electrical load centers of energy generation units. Virtual electric system is organized on the basis of cellular communication and central computer (smart grid). The algorithm is realized in the computer-aided design MathCAD. Using this program, the calculation of the electric system of an industrial microdistrict of the city with distributed generation was performed. As a result of the calculations, cartograms of electrical loads are constructed and the loss of electric energy in distribution lines is estimated. The calculation results show a significant reduction in losses in distribution lines when using distributed generation with dynamic redistribution of power sources.

The manuscript analyses the management of low and ultra-low-temperature district heating systems (DHS) coupled with centralised and decentralised heat pumps. Operative conditions are defined in order to satisfy the heating needs without overloading the electric grid. The results are achieved by dynamic simulations, based on a real DHS located in southern Switzerland. At the building level, the heating needs are estimated considering real data and simultaneous energy simulations. Two DHS configurations, alternatives to the existing one, are simulated and suitable parameters for the management of the DHS are selected. The global performance of the two DHS is evaluated by KPIs also including the flexibility and the impact on the electric peak due to heat pumps. The achieved results are discussed providing suggestions for the stakeholders involved in DHS management for an optimal matching of the electric grid and thermal networks towards a reduction of the peak power. The rule-based control strategies defined allow the expected electric peak shaving and load levelling, conversely, the yearly energy consumptions are lightly increased and have to be further investigated. The outcomes demonstrate a global better performance of the ultra-low temperature DHS in terms of response to the applied control strategies and of energy savings.

In this paper, the fracture problem of a functionally graded piezoelectric material strip (FGPM strip) containing a crack perpendicular to the interface between the FGPM strip and a homogeneous layer under an electric load is considered. Material properties are assumed to be exponentially dependent on the distance from the interface. The superposition technique is used to solve the governing equations. The stresses induced by the electric load in the un-cracked laminate are calculated, and the obtained normal stress is used as the crack surface tractions with opposite sign to formulate the mixed boundary value problem. By using the Fourier transforms, the electro-mechanical fracture problem is reduced to a singular integral equation, which is solved numerically. The stress intensity factors of the internal crack and the edge crack are computed and presented for the various values of the nonhomogeneous and geometric parameters.

Accurate short-term load forecasting can ensure the safe operation of the grid. Decomposing load data into smooth components by decomposition algorithms is a common approach to address data volatility. However, each component of the decomposition must be modeled separately for prediction, which leads to overly complex models. To solve this problem, a VMD-WSLSTM load prediction model based on Shapley values is proposed in this paper. First, the Shapley value is used to select the optimal set of special features, and then the VMD decomposition method is used to decompose the original load into several smooth components. Finally, WSLSTM is used to predict each component. Unlike the traditional LSTM model, WSLSTM can simplify the prediction model and extract common features among the components by sharing the parameters among the components. In order to verify the effectiveness of the proposed model, several control groups were used for experiments. The results show that the proposed method has higher prediction accuracy and training speed compared with traditional prediction methods.

The basis for achieving high quality of electricity are three components: high-quality power generation, uninterrupted transmission and distribution over reliable networks. In conditions of commensurability of the power of electric motors with the power supply and with a considerable extension of the cable network, algorithms for automated design of the power supply system are needed. This article describes an approach based on the application of the Dijkstra algorithm, which allows to reduce costs in the construction of the distribution network and provide the necessary voltage level at the motor terminals or minimum energy losses in the cable network. The main and arbitrary structure of a power supply network with an electric load in terms of the theory of the genetic algorithm is described, on the basis of which it is possible to create an effective means of determining rational configurations of a power supply network with an electric load.

Nowadays we observe rather rapid growth of energy accumulators market. There are prerequisites to their extensive application in Belarus. In spite of technology development problems pertaining to optimization of electric power and their operation under conditions of specific systems “generator – accumulator – consumer” (GAC) have not obtained proper consideration. At the same time tuning and optimization of the GAC system may provide competitive advantages to various accumulating systems because application of accumulator batteries in non-optimal charge – discharge conditions reduces its operating resource. Optimization of the GAC system may include utilization of hybrid accumulator systems together with heterogeneous chemical and mechanical accumulators, tuning of system controller parameters etc. Research papers present a great number of empirical and analytical methods for calculation of electric loads. These methods use the following parameters as initial data: time-averaged values of actual electric power consumption, averaged apartment electric loads, empirical and statistical form coefficients, coefficients of maximum electric load for a group of uniform consumers. However such models do not meet the requirements of detailed simulation of relatively small system operation when the simulation must correspond to non-stationary, non-averaged, stochastic load nature. The paper provides a simple approach to the detailed simulation of electric loads in regard to small projects such as multi-unit apartment building or small agricultural farm. The model is formulated both in physical and algorithmic terms that make it possible to be easily realized in any programming environment. The paper presents convergence of integral electric power consumption, which is set by the model, to statistically averaged parameters. Autocorrelation function has been calculated in the paper that shows two scales for autocorrelation of simulated load diagrams. The paper contains of an analysis on Fourier spectra for generated electric load diagrams. The model of electric load may be used separately or as a component of a general model in the GAC system. It may also be applied for tuning of automatic systems for control and accounting of power resources and various research purposes.

Model of main structure of power-supply is considered for research of states of working collections of induction motors. Model consists of separate electromechanic modules, which can contain different amount of motors. Modules are connected in different points to main cable, laid from transformer before remoted module. On base of this model is developed software program, intended for finding of optimum parameters of structure of network of power-supply with use the genetic algorithm.

Electric load forecasting plays an important role in electricity markets and power systems. Because electric load time series are complicated and nonlinear, it is very difficult to achieve a satisfactory forecasting accuracy. In this paper, a hybrid model, Wavelet Denoising-Extreme Learning Machine optimized by k-Nearest Neighbor Regression (EWKM), which combines k-Nearest Neighbor (KNN) and Extreme Learning Machine (ELM) based on a wavelet denoising technique is proposed for short-term load forecasting. The proposed hybrid model decomposes the time series into a low frequency-associated main signal and some detailed signals associated with high frequencies at first, then uses KNN to determine the independent and dependent variables from the low-frequency signal. Finally, the ELM is used to get the non-linear relationship between these variables to get the final prediction result for the electric load. Compared with three other models, Extreme Learning Machine optimized by k-Nearest Neighbor Regression (EKM), Wavelet Denoising-Extreme Learning Machine (WKM) and Wavelet Denoising-Back Propagation Neural Network optimized by k-Nearest Neighbor Regression (WNNM), the model proposed in this paper can improve the accuracy efficiently. New South Wales is the economic powerhouse of Australia, so we use the proposed model to predict electric demand for that region. The accurate prediction has a significant meaning.

A Neural Network based full-week hourly electric load forecasting model is proposed for Colombia. This model uses historical information delays as well as previously identified date events which produce significant changes in the electric load patrons through the year, the model also consider a three weeks delay in the available information used in forecasts. The model was validated using real electric load data from a specific Colombian region. The results were compared with an auto-regressive model (AR) and an auto-regressive model with exogenous variables (ARX). The general error decay and the good approximation during the atypical time periods, which are difficult to forecast, make it a satisfying model.