A multi-objective predictive energy management strategy for residential grid-connected PV-battery hybrid systems based on machine learning technique

Optimal planning of hybrid energy systems has always been a considerable task. While several studies have conducted techno‐economic analysis on hybrid energy systems, most of them have neglected to consider appropriate electric and thermal load growth rates. This paper proposes an efficient strategy for optimizing a grid‐connected hybrid energy system considering electric and thermal load growth rates. HOMER software is used for optimizing hybrid energy systems. It uses a nonderivative optimization to recognize the system with the minimum net present cost (NPC) among hundreds of configurations. The exponential smoothing method is also used for predicting upcoming load peak values based on historical data. The proposed hybrid system includes solar photovoltaic (SPV) system, wind turbine (WT) system, converter, battery storage system (BSS), gas generator (GasGen), fuel cell (FC), fuel‐fired boiler, and resistive boiler. Unlike previous studies, electricity price growth and SPV degradation rates are considered in this study. The grid‐connected system has also been compared to the stand‐alone system to understand grid connection impacts on the optimization results. Environmental analysis has been performed to analyze the greenhouse gas emissions of each system. The effects of growth rates in thermal and electric loads and electricity price are comprehensively investigated. Neglecting load growth rates affected the financial results severely. In this condition, the NPC and COE are achieved −11 383.4 US$ and −0.02376 US$/kWh. Results indicate that the combination of WT system (15 kW), converter (10 kW), resistive, and fuel‐fired (10 kW) boilers is the most economical configuration in grid‐connected mode. The WT system plays an essential role in the electric power provision and revenues of the hybrid system. The net present cost (NPC) and the cost of energy (COE) for grid‐connected operating mode are achieved −0.0162272 US$/kWh and −5243.956 US$, respectively. For off‐grid operating mode, the NPC and COE are obtained 47 024.19 US$ and 0.1983879 US$/kWh, respectively. Thanks to the feed‐in‐tariff (FiT) policy, the proposed grid‐connected system is more economical and environmentally friendly than the stand‐alone system. In off‐grid operating mode, 63 575.45285 kg CO2 is produced, which is higher than CO2 production in the grid‐connected mode (more than 8262 kg).

Today’s remarkable challenge of maritime transportation industry is the detrimental contamination generation from fossil fuels. To tackle such a challenge and reduce the contribution into air pollution, different power solutions have been considered; among others, hybrid energy-based solutions are powering many ferry boats. This paper introduces an energy management strategy (EMS) for a hybrid energy system (HES) of a ferry boat with the goal to optimize the performance and reduce the operation cost. HES considered for the ferry boat consists of different devices such as proton exchange membrane fuel cell (PEMFC), LI-ION battery bank, and cold ironing (CI). PEMFC systems are appropriate to employ as they are not polluting. The battery bank compensates for the abrupt variations of the load as the fuel cell has a slow dynamic against sudden changes of the load. Also, CI systems can improve the reduction of the expenses of energy management, during hours where the ferry boat is located at the harbor. To study the performance, cost and the pollution contribution CO2, NOX, SOX of the proposed hybrid energy management strategy (HEMS), we compare it against three various types of HEM from the state-of-the-art and also available rule-based methods in the literature. The analysis results show a high applicability of the proposed HES. All results in this paper have been obtained in the MATLAB software environment.

Short-Term Electricity Load Forecasting (STELF) through Data Analytics (DA) is an emerging and active research area. Forecasting about electricity load and price provides future trends and patterns of consumption. There is a loss in generation and use of electricity. So, multiple strategies are used to solve the aforementioned problems. Day-ahead electricity price and load forecasting are beneficial for both suppliers and consumers. In this paper, Deep Learning (DL) and data mining techniques are used for electricity load and price forecasting. XG-Boost (XGB), Decision Tree (DT), Recursive Feature Elimination (RFE) and Random Forest (RF) are used for feature selection and feature extraction. Enhanced Convolutional Neural Network (ECNN) and Enhanced Support Vector Regression (ESVR) are used as classifiers. Grid Search (GS) is used for tuning of the parameters of classifiers to increase their performance. The risk of over-fitting is mitigated by adding multiple layers in ECNN. Finally, the proposed models are compared with different benchmark schemes for stability analysis. The performance metrics MSE, RMSE, MAE, and MAPE are used to evaluate the performance of the proposed models. The experimental results show that the proposed models outperformed other benchmark schemes. ECNN performed well with threshold 0.08 for load forecasting. While ESVR performed better with threshold value 0.15 for price forecasting. ECNN achieved almost 2% better accuracy than CNN. Furthermore, ESVR achieved almost 1% better accuracy than the existing scheme (SVR).

Hybrid energy systems with renewable generation and battery energy storage system are becoming more popular in last decades due to the trend of decarbonization. The efficient functioning of such systems requires accurate forecasting of generation and consumption. It allows increasing energy efficiency by reducing the payment for energy supply and increasing the installed capacity factor of renewable generation. In this paper, we consider the use of various methods for forecasting electrical consumption and solar generation in hybrid energy systems. The authors implemented 11 forecasting methods with the use of python libraries on the example of residential building with high share of solar generation. To improve accuracy, we performed the feature importance analysis and evaluated the forecasting results. It allowed selecting the best methods for load and solar generation forecasting. The obtained results can be used for improving the efficiency of battery energy storage system control algorithms. It will ensure the maximum reduction in electricity bills and power consumption while increasing capacity utilization factor of solar power plant.

This paper proposes a model having both linear and nonlinear system dynamics by integrating both autoregressive integrated moving average (ARIMA) model and artificial neural network (ANN) model to simulate electrical energy supply inherent with strong seasonal and periodic characteristics in power system. Accurate electrical load forecast becomes possible by the integrated model for the ARIMA is effective to electricity load time series inherent with seasonal fluctuations as well as strong 7-day (per week) periodic characteristics. By using the input of historical daily electricity load data, weather data, and holiday effect variables, the integrated model is shown to be more accurate than the ANN model, the ARIMA model, the classical ARIMA-ANN model, and other well-known methods in the prediction and the forecast of electrical load in normal summer week, normal winter week, 3/4-day holiday week, long holiday week, and extreme weather week.

The powertrain model of the series-parallel plug-in hybrid electric vehicles (PHEVs) is more complicated, compared with series PHEVs and parallel PHEVs. Using the traditional dynamic programming (DP) algorithm or Pontryagin minimum principle (PMP) algorithm to solve the global-optimization-based energy management strategies of the series-parallel PHEVs is not ideal, as the solution time is too long or even impossible to solve. Chief engineers of hybrid system urgently require a handy tool to quickly solve global-optimization-based energy management strategies. Therefore, this paper proposed to use the Radau pseudospectral knotting method (RPKM) to solve the global-optimization-based energy management strategy of the series-parallel PHEVs to improve computational efficiency. Simulation results showed that compared with the DP algorithm, the global-optimization-based energy management strategy based on the RPKM improves the computational efficiency by 1806 times with a relative error of only 0.12%. On this basis, a bi-level nested component-sizing method combining the genetic algorithm and RPKM was developed. By applying the global-optimization-based energy management strategy based on RPKM to the actual development, the feasibility and superiority of RPKM applied to the global-optimization-based energy management strategy of the series-parallel PHEVs were further verified.

RETRACTED: Risk-constrained optimal operation of fuel cell/photovoltaic/battery/grid hybrid energy system using downside risk constraints method

This paper presents optimization of hybrid energy system for electrification of Rajasthan Technical University campus located at Kota, Rajasthan in India. The hybrid energy system comprises more than one energy sources such as solar PV-Diesel Generator or solar photo voltaic (PV)-wind turbine (WT)-Diesel Hybrid system. The main reason for hybrid energy system development is the reliability of solar and wind energy system decreases due to intermittent nature of these resources and availability. The hybrid system can efficiently utilize the resources as solar, wind, biomass, and transmission and distribution system. Electric load estimation is done according to IEC standards for institute. The component chosen for system is solar PV, small wind turbine, battery bank. Various sensitivity analysis and simulations are done to optimize the system using Optimization tool Hybrid Optimization Model for Electric Renewable (HOMER Pro). Fuel cell system did not use in the study due to non-availability of fuel, higher cost, and lower life time. The HOMER optimized the system in constrained of various costs of system, percentage of renewable energy uses, carbon emission, and electrical load requirement throughout the year.

An integrated framework for sizing and energy management of hybrid energy systems using finite automata

Energy management of hybrid energy system sources based on machine learning classification algorithms

An energy management strategy for plug-in hybrid electric vehicles based on deep learning and improved model predictive control

The energy management strategy renders significant impacts on the operating performance of plug-in hybrid electric vehicles (PHEV), which are similar to the traditional hybrid electric vehicle. A comprehensive methodology based on Particle Swarm Optimization (PSO) is presented in the paper to achieve parameter optimization for the energy management strategy, with a view to reducing the fuel consumption of PHEV. The parameters of energy management strategy are set as the optimized variables by PSO, with dynamic performance index of PHEV being defined as the constraint condition. Computer simulations are hence carried out, which show the PSO scheme gives much preferable results to original energy management strategy, and thereby the fuel consumption of PHEV can be effectively reduced without sacrificing PHEV's dynamic performance.

Energy management strategy of hybrid energy system for a multi-lobes hybrid air vehicle

Short-term load forecasting is an essential instrument in power system planning, operation, and control. Many operating decisions are based on load forecasts, such as dispatch scheduling of generating capacity, reliability analysis, and maintenance planning for the generators. This paper discusses significant role of artificial intelligence (AI) in short-term load forecasting (STLF), that is, the day-ahead hourly forecast of the power system load. A new artificial neural network (ANN) has been designed to compute the forecasted load. The data used in the modeling of ANN are hourly historical data of the temperature and electricity load. The ANN model is trained on hourly data from Ontario Electricity Market from 2007 to 2011 and tested on out-of-sample data from 2012. Simulation results obtained have shown that day-ahead hourly forecasts of load using proposed ANN is very accurate with very less error. However load forecast considering the effect of temperature is better than without taking it as input parameter.

Day ahead hourly load forecasting is an essential instrument in power system planning, operation, and control. Many operating decisions are based on load forecasts, such as dispatch scheduling of generating capacity, reliability analysis, and maintenance planning for the generators. This paper discusses significant role of artificial intelligence (AI) in short-term load forecasting (STLF), that is, the day-ahead hourly forecast of the power system load. A new artificial neural network (ANN) has been designed to compute the forecasted load. The data used in the modeling of ANN are hourly historical data of the temperature and electricity load. The ANN model is trained on hourly data from the ISO New England market and PJM Electricity Market from 2007 to 2011 and tested on out-of-sample data from 2012. Simulation results obtained have shown that day-ahead hourly forecasts of load using proposed ANN is very accurate with very less error in both the markets. However load forecast for ISO New England market is better than PJM market as temperature data has also been considered as input to ANN for this market.